Compare commits

...

293 Commits

Author SHA1 Message Date
License Updater
5ded954461 licenses: update android licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-03-27 17:40:25 +00:00
Patrick O'Doherty
8f27520633 safeweb: init (#11467)
Updates https://github.com/tailscale/corp/issues/8027

Safeweb is a wrapper around http.Server & tsnet that encodes some
application security defaults.

Safeweb asks developers to split their HTTP routes into two
http.ServeMuxs for serving browser and API-facing endpoints
repsectively. It then wraps these HTTP routes with the
context-appropriate security controls.

safeweb.Server#Serve will serve the HTTP muxes over the provided
listener. Caller are responsible for creating and tearing down their
application's listeners. Applications being served over HTTPS that wish
to implement HTTP redirects can use the Server#HTTPRedirect handler to
do so.

Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
2024-03-27 10:10:59 -07:00
Andrea Gottardo
008676f76e cmd/serve: update warning for sandboxed macOS builds (#11530) 2024-03-27 09:03:52 -07:00
Percy Wegmann
66e4d843c1 ipn/localapi: add support for multipart POST to file-put
This allows sending multiple files via Taildrop in one request.
Progress is tracked via ipn.Notify.

Updates tailscale/corp#18202

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-27 08:53:52 -05:00
Percy Wegmann
bed818a978 ipn/localapi: add support for multipart POST to file-put
This allows sending multiple files via Taildrop in one request.
Progress is tracked via ipn.Notify.

Updates tailscale/corp#18202

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-27 08:53:52 -05:00
Maisem Ali
0d8cd1645a go.mod: bump github.com/gaissmai/bart
To pick up https://github.com/gaissmai/bart/pull/17.

Updates #deps

Signed-off-by: Maisem Ali <maisem@tailscale.com>
2024-03-26 16:06:52 -07:00
Percy Wegmann
eb42a16da9 ipn/ipnlocal: report Taildrive access message on failed responses
For example, if we get a 404 when downloading a file, we'll report access.

Also, to reduce verbosty of logs, this elides 0 length files.

Updates tailscale/corp#17818

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-26 16:37:08 -05:00
Andrew Lytvynov
5d41259a63 cmd/tailscale/cli: remove Beta tag from tailscale update (#11529)
Updates #cleanup

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-03-26 15:28:34 -06:00
Charlotte Brandhorst-Satzkorn
acb611f034 ipn/localipn: introduce logs for tailfs (#11496)
This change introduces some basic logging into the access and share
pathways for tailfs.

Updates tailscale/corp#17818

Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
2024-03-26 13:14:43 -07:00
Irbe Krumina
4cbef20569 cmd/k8s-operator: redact auth key from debug logs (#11523)
Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-03-26 16:20:32 +00:00
Brad Fitzpatrick
55baf9474f metrics, tsweb/varz: add multi-label map metrics
Updates tailscale/corp#18640

Change-Id: Ia9ae25956038e9d3266ea165537ac6f02485b74c
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-26 09:01:04 -07:00
Flakes Updater
90a4d6ce69 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-03-26 08:37:52 -07:00
Brad Fitzpatrick
6d90966c1f logtail: move a scratch buffer to Logger
Rather than pass around a scratch buffer, put it on the Logger.

This is a baby step towards removing the background uploading
goroutine and starting it as needed.

Updates tailscale/corp#18514 (insofar as it led me to look at this code)

Change-Id: I6fd94581c28bde40fdb9fca788eb9590bcedae1b
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-25 17:33:42 -07:00
Irbe Krumina
06e22a96b1 .github/workflows: fix path filter for 'Kubernetes manifests' test job (#11520)
Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-03-25 19:44:30 +00:00
Chris Milson-Tokunaga
b6dfd7443a Change type of installCRDs (#11478)
Including the double quotes (`"`) around the value made it appear like the helm chart should expect a string value for `installCRDs`.

Signed-off-by: Chris Milson-Tokunaga <chris.w.milson@gmail.com>
2024-03-25 19:11:55 +00:00
Percy Wegmann
8b8b315258 net/tstun: use gaissmai/bart instead of tempfork/device
This implementation uses less memory than tempfork/device,
which helps avoid OOM conditions in the iOS VPN extension when
switching to a Tailnet with ExitNode routing enabled.

Updates tailscale/corp#18514

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-25 12:31:14 -05:00
Andrew Lytvynov
1e7050e73a go.mod: bump github.com/docker/docker (#11515)
There's a vulnerability https://pkg.go.dev/vuln/GO-2024-2659 that
govulncheck flags, even though it's only reachable from tests and
cmd/sync-containers and cannot be exploited there.

Updates #cleanup

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-03-25 10:45:35 -06:00
Brad Fitzpatrick
a36cfb4d3d tailcfg, ipn/ipnlocal, wgengine/magicsock: add only-tcp-443 node attr
Updates tailscale/corp#17879

Change-Id: I0dc305d147b76c409cf729b599a94fa723aef0e0
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-25 08:48:25 -07:00
Brad Fitzpatrick
7b34154df2 all: deprecate Node.Capabilities (more), remove PeerChange.Capabilities [capver 89]
First we had Capabilities []string. Then
https://tailscale.com/blog/acl-grants (#4217) brought CapMap, a
superset of Capabilities. Except we never really finished the
transition inside the codebase to go all-in on CapMap. This does so.

Notably, this coverts Capabilities on the wire early to CapMap
internally so the code can only deal in CapMap, even against an old
control server.

In the process, this removes PeerChange.Capabilities support, which no
known control plane sent anyway. They can and should use
PeerChange.CapMap instead.

Updates #11508
Updates #4217

Change-Id: I872074e226b873f9a578d9603897b831d50b25d9
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-24 21:08:46 -07:00
Brad Fitzpatrick
4992aca6ec tsweb/varz: flesh out munging of expvar keys into valid Prometheus metrics
From a problem we hit with how badger registers expvars; it broke
trunkd's exported metrics.

Updates tailscale/corp#1297

Change-Id: I42e1552e25f734c6f521b6e993d57a82849464b2
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-24 20:06:06 -07:00
Brad Fitzpatrick
b104688e04 ipn/ipnlocal, types/netmap: replace hasCapability with set lookup on NetworkMap
When node attributes were super rare, the O(n) slice scans looking for
node attributes was more acceptable. But now more code and more users
are using increasingly more node attributes. Time to make it a map.

Noticed while working on tailscale/corp#17879

Updates #cleanup

Change-Id: Ic17c80341f418421002fbceb47490729048756d2
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-22 15:30:46 -07:00
Percy Wegmann
8c88853db6 ipn/ipnlocal: add c2n /debug/pprof/allocs endpoint
This behaves the same as typical debug/pprof/allocs.

Updates tailscale/corp#18514

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-22 17:29:59 -05:00
Brad Fitzpatrick
f45594d2c9 control/controlclient: free memory on iOS before full netmap work
Updates tailscale/corp#18514

Change-Id: I8d0330334b030ed8692b25549a0ee887ac6d7188
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-22 09:02:19 -07:00
James Tucker
e0f97738ee localapi: reduce garbage production in bus watcher
Updates #optimization

Signed-off-by: James Tucker <james@tailscale.com>
2024-03-21 17:50:20 -07:00
James Tucker
3f7313dbdb util/linuxfw,wgengine/router: enable IPv6 configuration when netfilter is disabled
Updates #11434

Signed-off-by: James Tucker <james@tailscale.com>
2024-03-21 16:10:47 -07:00
Brad Fitzpatrick
8444937c89 control/controlclient: fix panic regression from earlier load balancer hint header
In the recent 20e9f3369 we made HealthChangeRequest machine requests
include a NodeKey, as it was the oddball machine request that didn't
include one. Unfortunately, that code was sometimes being called (at
least in some of our integration tests) without a node key due to its
registration with health.RegisterWatcher(direct.ReportHealthChange).

Fortunately tests in corp caught this before we cut a release. It's
possible this only affects this particular integration test's
environment, but still worth fixing.

Updates tailscale/corp#1297

Change-Id: I84046779955105763dc1be5121c69fec3c138672
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-21 12:54:58 -07:00
Joe Tsai
85febda86d all: use zstdframe where sensible (#11491)
Use the zstdframe package where sensible instead of plumbing
around our own zstd.Encoder just for stateless operations.

This causes logtail to have a dependency on zstd,
but that's arguably okay since zstd support is implicit
to the protocol between a client and the logging service.
Also, virtually every caller to logger.NewLogger was
manually setting up a zstd.Encoder anyways,
meaning that zstd was functionally always a dependency.

Updates #cleanup
Updates tailscale/corp#18514

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-03-21 12:20:38 -07:00
Joe Tsai
d4bfe34ba7 util/zstdframe: add package for stateless zstd compression (#11481)
The Go zstd package is not friendly for stateless zstd compression.
Passing around multiple zstd.Encoder just for stateless compression
is a waste of memory since the memory is never freed and seldom
used if no compression operations are happening.

For performance, we pool the relevant Encoder/Decoder
with the specific options set.

Functionally, this package is a wrapper over the Go zstd package
with a more ergonomic API for stateless operations.

This package can be used to cleanup various pre-existing zstd.Encoder
pools or one-off handlers spread throughout our codebases.

Performance:

	BenchmarkEncode/Best               1690        610926 ns/op      25.78 MB/s           1 B/op          0 allocs/op
	    zstd_test.go:137: memory: 50.336 MiB
	    zstd_test.go:138: ratio:  3.269x
	BenchmarkEncode/Better            10000        100939 ns/op     156.04 MB/s           0 B/op          0 allocs/op
	    zstd_test.go:137: memory: 20.399 MiB
	    zstd_test.go:138: ratio:  3.131x
	BenchmarkEncode/Default            15775         74976 ns/op     210.08 MB/s         105 B/op          0 allocs/op
	    zstd_test.go:137: memory: 1.586 MiB
	    zstd_test.go:138: ratio:  3.064x
	BenchmarkEncode/Fastest            23222         53977 ns/op     291.81 MB/s          26 B/op          0 allocs/op
	    zstd_test.go:137: memory: 599.458 KiB
	    zstd_test.go:138: ratio:  2.898x
	BenchmarkEncode/FastestLowMemory                   23361         50789 ns/op     310.13 MB/s          15 B/op          0 allocs/op
	    zstd_test.go:137: memory: 334.458 KiB
	    zstd_test.go:138: ratio:  2.898x
	BenchmarkEncode/FastestNoChecksum                  23086         50253 ns/op     313.44 MB/s          26 B/op          0 allocs/op
	    zstd_test.go:137: memory: 599.458 KiB
	    zstd_test.go:138: ratio:  2.900x

	BenchmarkDecode/Checksum                           70794         17082 ns/op     300.96 MB/s           4 B/op          0 allocs/op
	    zstd_test.go:163: memory: 316.438 KiB
	BenchmarkDecode/NoChecksum                         74935         15990 ns/op     321.51 MB/s           4 B/op          0 allocs/op
	    zstd_test.go:163: memory: 316.438 KiB
	BenchmarkDecode/LowMemory                          71043         16739 ns/op     307.13 MB/s           0 B/op          0 allocs/op
	    zstd_test.go:163: memory: 79.347 KiB

We can see that the options are taking effect where compression ratio improves
with higher levels and compression speed diminishes.
We can also see that LowMemory takes effect where the pooled coder object
references less memory than other cases.
We can see that the pooling is taking effect as there are 0 amortized allocations.

Additional performance:

	BenchmarkEncodeParallel/zstd-24                     1857        619264 ns/op        1796 B/op         49 allocs/op
	BenchmarkEncodeParallel/zstdframe-24                1954        532023 ns/op        4293 B/op         49 allocs/op
	BenchmarkDecodeParallel/zstd-24                     5288        197281 ns/op        2516 B/op         49 allocs/op
	BenchmarkDecodeParallel/zstdframe-24                6441        196254 ns/op        2513 B/op         49 allocs/op

In concurrent usage, handling the pooling in this package
has a marginal benefit over the zstd package,
which relies on a Go channel as the pooling mechanism.
In particular, coders can be freed by the GC when not in use.
Coders can be shared throughout the program if they use this package
instead of multiple independent pools doing the same thing.
The allocations are unrelated to pooling as they're caused by the spawning of goroutines.

Updates #cleanup
Updates tailscale/corp#18514
Updates tailscale/corp#17653
Updates tailscale/corp#18005

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-03-21 11:39:20 -07:00
Brad Fitzpatrick
6a860cfb35 ipn/ipnlocal: add c2n pprof option to force a GC
Like net/http/pprof has.

Updates tailscale/corp#18514

Change-Id: I264adb6dcf5732d19707783b29b7273b4ca69cf4
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-21 10:24:06 -07:00
Brad Fitzpatrick
5d1c72f76b wgengine/magicsock: don't use endpoint debug ringbuffer on mobile.
Save some memory.

Updates tailscale/corp#18514

Change-Id: Ibcaf3c6d8e5cc275c81f04141d0f176e2249509b
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-21 06:58:55 -07:00
Andrew Dunham
512fc0b502 util/reload: add new package to handle periodic value loading
This can be used to reload a value periodically, whether from disk or
another source, while handling jitter and graceful shutdown.

Updates tailscale/corp#1297

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Iee2b4385c9abae59805f642a7308837877cb5b3f
2024-03-20 18:23:54 -04:00
Adrian Dewhurst
2f7e7be2ea control/controlclient: do not alias peer CapMap
Updates #cleanup

Change-Id: I10fd5e04310cdd7894a3caa3045b86eb0a06b6a0
Signed-off-by: Adrian Dewhurst <adrian@tailscale.com>
2024-03-20 15:42:44 -04:00
Percy Wegmann
067ed0bf6f ipnlocal: ensure TailFS share notifications are non-nil
This allows the UI to distinguish between 'no shares' versus
'not being notified about shares'.

Updates ENG-2843

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-20 13:54:33 -05:00
Brad Fitzpatrick
20e9f3369d control/controlclient: send load balancing hint HTTP request header
Updates tailscale/corp#1297

Change-Id: I0b102081e81dfc1261f4b05521ab248a2e4a1298
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-20 09:39:41 -07:00
Percy Wegmann
15c58cb77c tailfs: include whitespace in test share and filenames
Since TailFS allows spaces in folder and file names, test with spaces.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-20 09:41:18 -05:00
James Tucker
e37eded256 tool/gocross: add android autoflags (#11465)
Updates tailscale/corp#18202

Signed-off-by: James Tucker <james@tailscale.com>
2024-03-19 16:08:20 -07:00
Claire Wang
221de01745 control/controlclient: fix sending peer capmap changes (#11457)
Instead of just checking if a peer capmap is nil, compare the previous
state peer capmap with the new peer capmap.
Updates tailscale/corp#17516

Signed-off-by: Claire Wang <claire@tailscale.com>
2024-03-19 18:56:06 -04:00
Andrew Dunham
6da1dc84de wgengine: fix logger data race in tests
Observed in:
    https://github.com/tailscale/tailscale/actions/runs/8350904950/job/22858266932?pr=11463

Updates #11226

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I9b57db4b34b6ad91d240cd9fa7e344fc0376d52d
2024-03-19 18:51:25 -04:00
Andrew Dunham
e382e4cee6 syncs: add Swap method
To mimic sync.Map.Swap, sync/atomic.Value.Swap, etc.

Updates tailscale/corp#1297

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: If7627da1bce8b552873b21d7e5ebb98904e9a650
2024-03-19 18:39:44 -04:00
Andrea Gottardo
6288c9b41e version/prop: remove IsMacAppSandboxEnabled (#11461)
Fixes tailscale/corp#18441

For a few days, IsMacAppStore() has been returning `false` on App Store builds (IPN-macOS target in Xcode).

I regressed this in #11369 by introducing logic to detect the sandbox by checking for the APP_SANDBOX_CONTAINER_ID environment variable. I thought that was a more robust approach instead of checking the name of the executable. However, it appears that on recent macOS versions this environment variable is no longer getting set, so we should go back to the previous logic that checks for the executable path, or HOME containing references to macsys.

This PR also adds additional checks to the logic by also checking XPC_SERVICE_NAME in addition to HOME where possible. That environment variable is set inside the network extension, either macos or macsys and is good to look at if for any reason HOME is not set.
2024-03-19 14:50:34 -07:00
Mario Minardi
68d9e49a5b api.md: add missing backtick to GET searchpaths doc (#11459)
Add missing backtick to GET searchpaths api documentation.

Updates #cleanup

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-03-19 11:31:03 -06:00
Will Norris
349799a1ba api.md: format API docs with prettier
Mostly inconsequential minor fixes for consistency.  A couple of changes
to actual JSON examples, but all still very readable, so I think it's
fine.

Updates #cleanup

Signed-off-by: Will Norris <will@tailscale.com>
2024-03-19 09:23:50 -07:00
Irbe Krumina
b0c3e6f6c5 cmd/k8s-operator,ipn/conf.go: fix --accept-routes for proxies (#11453)
Fix a bug where all proxies got configured with --accept-routes set to true.
The bug was introduced in https://github.com/tailscale/tailscale/pull/11238.

Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-03-19 14:54:17 +00:00
James Tucker
7fe4cbbaf3 types/views: optimize slices contains under some conditions (#11449)
In control there are conditions where the leaf functions are not being
optimized away (i.e. At is not inlined), resulting in undesirable time
spent copying during SliceContains. This optimization is likely
irrelevant to simpler code or smaller structures.

Updates #optimization

Signed-off-by: James Tucker <james@tailscale.com>
2024-03-18 16:19:16 -07:00
Mario Minardi
d2ccfa4edd cmd/tailscale,ipn/ipnlocal: enable web client over quad 100 by default (#11419)
Enable the web client over 100.100.100.100 by default. Accepting traffic
from [tailnet IP]:5252 still requires setting the `webclient` user pref.

Updates https://github.com/tailscale/tailscale/issues/10261

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-03-18 15:47:21 -06:00
Will Norris
4d747c1833 api.md: document device expiration endpoint
This was originally built for testing node expiration flows, but is also
useful for customers to force device re-auth without actually deleting
the device from the tailnet.

Updates tailscale/corp#18408

Signed-off-by: Will Norris <will@tailscale.com>
2024-03-18 12:49:55 -07:00
Mario Minardi
e0886ad167 ipn/ipnlocal, tailcfg: add disable-web-client node attribute (#11418)
Add a disable-web-client node attribute and add handling for disabling
the web client when this node attribute is set.

Updates https://github.com/tailscale/tailscale/issues/10261

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-03-18 10:32:33 -06:00
Marwan Sulaiman
da7c3d1753 envknob: ensure f is not nil before using it
This PR fixes a panic that I saw in the mac app where
parsing the env file fails but we don't get to see the
error due to the panic of using f.Name()

Fixes #11425

Signed-off-by: Marwan Sulaiman <marwan@tailscale.com>
2024-03-15 12:46:41 -04:00
Andrea Gottardo
08ebac9acb version,cli,safesocket: detect non-sandboxed macOS GUI (#11369)
Updates ENG-2848

We can safely disable the App Sandbox for our macsys GUI, allowing us to use `tailscale ssh` and do a few other things that we've wanted to do for a while. This PR:

- allows Tailscale SSH to be used from the macsys GUI binary when called from a CLI
- tweaks the detection of client variants in prop.go, with new functions `IsMacSys()`, `IsMacSysApp()` and `IsMacAppSandboxEnabled()`

Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
2024-03-14 14:28:06 -07:00
Irbe Krumina
ea55f96310 cmd/tailscale/cli: fix configuring partially empty kubeconfig (#11417)
When a user deletes the last cluster/user/context from their
kubeconfig via 'kubectl delete-[cluster|user|context] command,
kubectx sets the relevant field in kubeconfig to 'null'.
This was breaking our conversion logic that was assuming that the field
is either non-existant or is an array.

Updates tailscale/corp#18320

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-03-14 20:26:20 +00:00
Anton Tolchanov
cf8948da5f net/routetable: increase route limit used by the test
I was running all tests while preparing a recent stable release, and
this was failing because my computer is connected to a fairly large
tailnet.

```
--- FAIL: TestGetRouteTable (0.01s)
    routetable_linux_test.go:32: expected at least one default route;
    ...
```

```
$ ip route show table 52  | wc -l
1051
```

Updates #cleanup

Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-03-14 16:10:40 +00:00
Andrew Lytvynov
decd9893e4 ipn/ipnlocal: validate domain of PopBrowserURL on default control URL (#11394)
If the client uses the default Tailscale control URL, validate that all
PopBrowserURLs are under tailscale.com or *.tailscale.com. This reduces
the risk of a compromised control plane opening phishing pages for
example.

The client trusts control for many other things, but this is one easy
way to reduce that trust a bit.

Fixes #11393

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-03-13 18:31:07 -06:00
Andrew Lytvynov
48eef9e6eb clientupdate: do not allow msiexec to reboot the OS (#11409)
According to
https://learn.microsoft.com/en-us/windows/win32/msi/standard-installer-command-line-options#promptrestart,
`/promptrestart` is ignored with `/quiet` is set, so msiexec.exe can
sometimes silently trigger a reboot. The best we can do to reduce
unexpected disruption is to just prevent restarts, until the user
chooses to do it. Restarts aren't normally needed for Tailscale updates,
but there seem to be some situations where it's triggered.

Updates #18254

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-03-13 16:55:24 -06:00
Anton Tolchanov
da3cf12194 VERSION.txt: this is v1.63.0
Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-03-13 14:51:52 +00:00
Anton Tolchanov
f12d2557f9 prober: add a DERP bandwidth probe
Updates tailscale/corp#17912

Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-03-13 13:36:45 +00:00
Anton Tolchanov
5018683d58 prober: remove unused derp prober latency measurements
Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-03-13 13:36:45 +00:00
Anton Tolchanov
205a10b51a prober: export probe counters and cumulative latency
Updates #cleanup

Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-03-13 13:36:45 +00:00
Andrew Dunham
7429e8912a wgengine/netstack: fix bug with duplicate SYN packets in client limit
This fixes a bug that was introduced in #11258 where the handling of the
per-client limit didn't properly account for the fact that the gVisor
TCP forwarder will return 'true' to indicate that it's handled a
duplicate SYN packet, but not launch the handler goroutine.

In such a case, we neither decremented our per-client limit in the
wrapper function, nor did we do so in the handler function, leading to
our per-client limit table slowly filling up without bound.

Fix this by doing the same duplicate-tracking logic that the TCP
forwarder does so we can detect such cases and appropriately decrement
our in-flight counter.

Updates tailscale/corp#12184

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ib6011a71d382a10d68c0802593f34b8153d06892
2024-03-11 08:05:00 -04:00
Brad Fitzpatrick
ad33e47270 ipn/{ipnlocal,localapi}: add debug verb to force spam IPN bus NetMap
To force the problem in its worst case scenario before fixing it.

Updates tailscale/corp#17859

Change-Id: I2c8b8e5f15c7801e1ab093feeafac52ec175a763
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-09 17:47:15 -08:00
Flakes Updater
04fceae898 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-03-09 11:51:43 -08:00
James Tucker
055117ad45 util/linuxfw: fix support for containers without IPv6 iptables filters (#11381)
There are container environments such as GitHub codespaces that have
partial IPv6 support - routing support is enabled at the kernel level,
but lacking IPv6 filter support in the iptables module.

In the specific example of the codespaces environment, this also has
pre-existing legacy iptables rules in the IPv4 tables, as such the
nascent firewall mode detection will always pick iptables.

We would previously fault trying to install rules to the filter table,
this catches that condition earlier, and disables IPv6 support under
these conditions.

Updates #5621
Updates #11344
Updates #11354

Signed-off-by: James Tucker <james@tailscale.com>
2024-03-08 15:46:21 -08:00
James Tucker
43fba6e04d util/linuxfw: correct logical error in NAT table check (#11380)
Updates #11344
Updates #11354

Signed-off-by: James Tucker <james@tailscale.com>
2024-03-08 15:35:13 -08:00
panchajanya
50a570a83f Code Improvements (#11311)
build_docker, update-flake: cleanup and apply shellcheck fixes

Was editing this file to match my needs while shellcheck warnings
bugged me out.
REV isn't getting used anywhere. Better remove it.

Updates #cleanup

Signed-off-by: Panchajanya1999 <kernel@panchajanya.dev>
Signed-off-by: James Tucker <james@tailscale.com>
2024-03-08 15:24:36 -08:00
Percy Wegmann
e496451928 ipn,cmd/tailscale,client/tailscale: add support for renaming TailFS shares
- Updates API to support renaming TailFS shares.
- Adds a CLI rename subcommand for renaming a share.
- Renames the CLI subcommand 'add' to 'set' to make it clear that
  this is an add or update.
- Adds a unit test for TailFS in ipnlocal

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-08 14:48:26 -06:00
Percy Wegmann
6c160e6321 ipn,tailfs: tie TailFS share configuration to user profile
Previously, the configuration of which folders to share persisted across
profile changes. Now, it is tied to the user's profile.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-08 14:48:26 -06:00
Percy Wegmann
16ae0f65c0 cmd/viewer: import views when generating byteSliceField
Updates #cleanup

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-08 14:48:26 -06:00
Andrew Dunham
f072d017bd wgengine/magicsock: don't change DERP home when not connected to control
This pretty much always results in an outage because peers won't
discover our new home region and thus won't be able to establish
connectivity.

Updates tailscale/corp#18095

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ic0d09133f198b528dd40c6383b16d7663d9d37a7
2024-03-08 14:15:13 -05:00
Sonia Appasamy
54e52532eb version/mkversion: enforce synology versions within int32 range
Synology requires version numbers are within int32 range. This
change updates the version logic to keep things closer within the
range, and errors on building when the range is exceeded.

Updates #cleanup

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-03-08 12:47:59 -05:00
Claire Wang
74e33b9c50 tailcfg: bump CapabilityVersion (#11368)
bump version for adding NodeAttrSuggestExitNode
remove extra s from NodeAttrSuggestExitNode
Updates tailscale/corp#17516

Signed-off-by: Claire Wang <claire@tailscale.com>
2024-03-07 14:17:40 -05:00
Mario Minardi
c662bd9fe7 client/web: dedupe packages in yarn.lock (#11327)
Run yarn-deduplicate on yarn.lock to dedupe packages. This is being done
to reduce the number of redundant packages fetched by yarn when existing
versions in the lockfile satisfy the version dependency we need.

See https://github.com/scinos/yarn-deduplicate for details on the tool
used to perform this deduplication.

Updates #cleanup

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-03-07 09:29:20 -07:00
Andrew Dunham
34176432d6 cmd/derper, types/logger: move log filter to shared package
So we can use it in trunkd to quiet down the logs there.

Updates #5563

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ie3177dc33f5ad103db832aab5a3e0e4f128f973f
2024-03-07 11:05:03 -05:00
Irbe Krumina
3047b6274c docs/k8s: don't run subnet router in userspace mode (#11363)
There should not be a need to do that unless we run on host network

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-03-07 13:56:11 +00:00
Andrew Dunham
9884d06b80 net/interfaces: fix test hang on Darwin
This test could hang because the subprocess was blocked on writing to
the stdout pipe if we find the address we're looking for early in the
output.

Updates #cleanup

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I68d82c22a5d782098187ae6d8577e43063b72573
2024-03-06 22:37:40 -05:00
Andrew Dunham
62cf83eb92 go.mod: bump gvisor
The `stack.PacketBufferPtr` type no longer exists; replace it with
`*stack.PacketBuffer` instead.

Updates #8043

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ib56ceff09166a042aa3d9b80f50b2aa2d34b3683
2024-03-06 20:22:20 -05:00
Andrew Dunham
8f27d519bb tsweb: add String method to tsweb.RequestID
In case we want to change the format to something opaque later.

Updates tailscale/corp#2549

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ie2eac8b885b694be607e9d5101d24b650026d89c
2024-03-06 19:48:04 -05:00
Irbe Krumina
90c4067010 util/linuxfw: add container-friendly IPv6 NAT check (#11353)
Remove IPv6 NAT check when routing is being set up
using nftables.
This is unnecessary as support for nftables was
added after support for IPv6.
https://tldp.org/HOWTO/Linux+IPv6-HOWTO/ch18s04.html
https://wiki.nftables.org/wiki-nftables/index.php/Building_and_installing_nftables_from_sources

Additionally, run an extra check for IPv6 NAT support
when the routing is set up with iptables.
This is because the earlier checks rely on
being able to use modprobe and on /proc/net/ip6_tables_names
being populated on start - these conditions are usually not
true in container environments.

Updates tailscale/tailscale#11344

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-03-06 21:53:51 +00:00
Percy Wegmann
fd942b5384 ipn/ipnlocal: reduce allocations in TailFS share notifications
This eliminates unnecessary map.Clone() calls and also eliminates
repetitive notifications about the same set of shares.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-06 14:21:53 -06:00
Percy Wegmann
6f66f5a75a ipn: add comment about thread-safety to StateStore
Updates #cleanup

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-06 12:42:18 -06:00
Andrea Gottardo
0cb86468ca ipn/localapi: add set-gui-visible endpoint
Updates tailscale/corp#17859

Provides a local API endpoint to be called from the GUI to inform the backend when the client menu is opened or closed.

cc @bradfitz

Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
Signed-off-by: Andrea Gottardo <andrea@tailscale.com>
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
2024-03-06 10:39:52 -08:00
Percy Wegmann
00373f07ac ipn/ipnlocal: exclude mullvad exit nodes from TailFS peers list
This is a temporary solution to at least omit Mullvad exit nodes
from the list of TailFS peers. Once we can identify peers that are
actually sharing via TailFS, we can remove this, but for alpha it'll
be sufficient to just omit Mullvad.

Updates tailscale/corp#17766

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-06 12:27:32 -06:00
Sonia Appasamy
c58c59ee54 {ipn,cmd/tailscale/cli}: move ServeConfig mutation logic to ipn/serve
Moving logic that manipulates a ServeConfig into recievers on the
ServeConfig in the ipn package. This is setup work to allow the
web client and cli to both utilize these shared functions to edit
the serve config.

Any logic specific to flag parsing or validation is left untouched
in the cli command. The web client will similarly manage its
validation of user's requested changes. If validation logic becomes
similar-enough, we can make a serve util for shared functionality,
which likely does not make sense in ipn.

Updates #10261

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-03-06 13:26:03 -05:00
Kristoffer Dalby
65255b060b client/tailscale: add postures to UserRuleMatch
Updates tailscale/corp#17770

Signed-off-by: Kristoffer Dalby <kristoffer@tailscale.com>
2024-03-06 15:36:17 +01:00
License Updater
d59878e457 licenses: update android licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-03-05 17:55:26 -08:00
License Updater
797d75c50a licenses: update win/apple licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-03-05 17:53:52 -08:00
License Updater
6a4e5329c3 licenses: update tailscale{,d} licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-03-05 17:53:05 -08:00
Andrew Dunham
4338db28f7 wgengine/magicsock: prefer link-local addresses to private ones
Since link-local addresses are definitionally more likely to be a direct
(lower-latency, more reliable) connection than a non-link-local private
address, give those a bit of a boost when selecting endpoints.

Updates #8097

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I93fdeb07de55ba39ba5fcee0834b579ca05c2a4e
2024-03-05 20:32:45 -05:00
Sonia Appasamy
65c3c690cf {ipn/serve,cmd/tailscale/cli}: move some shared funcs to ipn
In preparation for changes to allow configuration of serve/funnel
from the web client, this commit moves some functionality that will
be shared between the CLI and web client to the ipn package's
serve.go file, where some other util funcs are already defined.

Updates #10261

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-03-05 14:30:38 -05:00
Brad Fitzpatrick
8780e33500 go.toolchain.rev: bump Go toolchain to 1.22.1
Updates tailscale/corp#18000

Change-Id: I45de95e974ea55b0dac2218b3c82d124c4793390
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-03-05 10:51:13 -08:00
Paul Scott
2fa20e3787 util/cmpver: add Less/LessEq helper funcs
Updates tailscale/corp#17199

Signed-off-by: Paul Scott <paul@tailscale.com>
2024-03-05 16:57:04 +00:00
Claire Wang
d610f8eec0 tailcfg: add suggest exit node related node attribute (#11329)
Updates tailscale/corp#17516

Signed-off-by: Claire Wang <claire@tailscale.com>
2024-03-05 10:54:41 -05:00
Chris Palmer
13853e7f29 tsweb: add more test cases for TestCleanRedirectURL (#11331)
Updates #cleanup

Signed-off-by: Chris Palmer <cpalmer@tailscale.com>
2024-03-04 17:13:36 -08:00
Irbe Krumina
dff6f3377f docs/k8s: update docs (#11307)
Update docs for static Tailscale deployments on kube
to always use firewall mode autodection when in non-userspace.
Also add a note about running multiple replicas and a few suggestions how folks could do that.

Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
Co-authored-by: Anton Tolchanov <1687799+knyar@users.noreply.github.com>
2024-03-04 14:59:51 +00:00
Percy Wegmann
232a2d627c tailfs: only impersonate unprivileged user if able to sudo -u as that user
When serving TailFS shares, tailscaled executes another tailscaled to act as a
file server. It attempts to execute this child process as an unprivileged user
using sudo -u. This is important to avoid accessing files as root, which would
result in potential privilege escalation.

Previously, tailscaled assumed that it was running as someone who can sudo -u,
and would fail if it was unable to sudo -u.

With this commit, if tailscaled is unable to sudo -u as the requested user, and
tailscaled is not running as root, then tailscaled executes the the file server
process under the same identity that ran tailscaled, since this is already an
unprivileged identity.

In the unlikely event that tailscaled is running as root but is unable to
sudo -u, it will refuse to run the child file server process in order to avoid
privilege escalation.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-03-01 17:42:19 -06:00
Flakes Updater
00554ad277 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-02-29 19:53:19 -08:00
Andrew Lytvynov
23fbf0003f clientupdate: handle multiple versions in "apk info tailscale" output (#11310)
The package info output can list multiple package versions, and not in
descending order. Find the newest version in the output, instead of the
first one.

Fixes #11309

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-02-29 11:54:46 -07:00
Irbe Krumina
097c5ed927 util/linuxfw: insert rather than append nftables DNAT rule (#11303)
Ensure that the latest DNATNonTailscaleTraffic rule
gets inserted on top of any pre-existing rules.

Updates tailscale/tailscale#11281

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-02-29 16:53:43 +00:00
Percy Wegmann
e324a5660f ipn: include full tailfs shares in ipn notifications
This allows the Mac application to regain access to restricted
folders after restarts.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-29 10:16:44 -06:00
Percy Wegmann
80f1cb6227 tailfs: support storing bookmark data on shares
This allows the sandboxed Mac application to store security-
scoped URL bookmarks in order to maintain access to restricted
folders across restarts.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-29 10:16:44 -06:00
Brad Fitzpatrick
f18f591bc6 wgengine: plumb the PeerByKey from wgengine to magicsock
This was just added in 69f4b459 which doesn't yet use it. This still
doesn't yet use it. It just pushes it down deeper into magicsock where
it'll used later.

Updates #7617

Change-Id: If2f8fd380af150ffc763489e1ff4f8ca2899fac6
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-28 19:36:34 -08:00
Andrew Lytvynov
c7474431f1 tsweb: allow empty redirect URL in CleanRedirectURL (#11295)
Updates #cleanup

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-02-28 15:57:42 -08:00
Brad Fitzpatrick
b68a09cb34 ipn/ipnlocal: make active IPN sessions keyed by sessionID
We used a HandleSet before when we didn't have a unique handle. But a
sessionID is a unique handle, so use that instead. Then that replaces
the other map we had.

And now we'll have a way to look up an IPN session by sessionID for
later.

Updates tailscale/corp#17859

Change-Id: I5f647f367563ec8783c643e49f93817b341d9064
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-28 15:16:10 -08:00
Percy Wegmann
2d5d6f5403 ipn,wgengine: only intercept TailFS traffic on quad 100
This fixes a regression introduced with 993acf4 and released in
v1.60.0.

The regression caused us to intercept all userspace traffic to port
8080 which prevented users from exposing their own services to their
tailnet at port 8080.

Now, we only intercept traffic to port 8080 if it's bound for
100.100.100.100 or fd7a:115c:a1e0::53.

Fixes #11283

Signed-off-by: Percy Wegmann <percy@tailscale.com>
(cherry picked from commit 17cd0626f3)
2024-02-28 17:09:14 -06:00
Ross Zurowski
e83e2e881b client/web: fix Vite CJS deprecation warning (#11288)
Starting in Vite 5, Vite now issues a deprecation warning when using
a CJS-based Vite config file. This commit fixes it by adding the
`"type": "module"` to our package.json to opt our files into ESM module
behaviours.

Fixes #cleanup

Signed-off-by: Ross Zurowski <ross@rosszurowski.com>
2024-02-28 16:28:22 -05:00
Brad Fitzpatrick
69f4b4595a wgengine{,/wgint}: add wgint.Peer wrapper type, add to wgengine.Engine
This adds a method to wgengine.Engine and plumbed down into magicsock
to add a way to get a type-safe Tailscale-safe wrapper around a
wireguard-go device.Peer that only exposes methods that are safe for
Tailscale to use internally.

It also removes HandshakeAttempts from PeerStatusLite that was just
added as it wasn't needed yet and is now accessible ala cart as needed
from the Peer type accessor.

None of this is used yet.

Updates #7617

Change-Id: I07be0c4e6679883e6eeddf8dbed7394c9e79c5f4
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-28 09:50:18 -08:00
James Tucker
7e17aeb36b .github/workflows: fix regular breakage of go toolchains
This server recently had a common ansible applied, which added a
periodic /tmp cleaner, as is needed on other CI machines to deal with
test tempfile leakage. The setting of $HOME to /tmp means that the go
toolchain in there was regularly getting pruned by the tmp cleaner, but
often incompletely, because it was also in use.

Move HOME to a runner owned directory.

Updates #11248

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-28 08:17:35 -08:00
Brad Fitzpatrick
b4ff9a578f wgengine: rename local variable from 'found' to conventional 'ok'
Updates #cleanup

Change-Id: I799dc86ea9e4a3a949592abdd8e74282e7e5d086
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-28 07:33:57 -08:00
Brad Fitzpatrick
a8a525282c wgengine: use slices.Clone in two places
Updates #cleanup

Change-Id: I1cb30efb6d09180e82b807d6146f37897ef99307
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-28 07:33:57 -08:00
Brad Fitzpatrick
74b8985e19 ipn/ipnstate, wgengine: make PeerStatusLite.LastHandshake zero Time means none
... rather than 1970. Code was using IsZero against the 1970 team
(which isn't a zero value), but fortunately not anywhere that seems to
have mattered.

Updates #cleanup

Change-Id: I708a3f2a9398aaaedc9503678b4a8a311e0e019e
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-28 07:33:57 -08:00
Andrew Dunham
3dd8ae2f26 net/tstun: fix spelling of "WireGuard"
Updates #cleanup

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ida7e30f4689bc18f5f7502f53a0adb5ac3c7981a
2024-02-28 00:00:18 -05:00
Andrew Dunham
a20e46a80f util/cache: fix missing interface methods (#11275)
Updates #cleanup


Change-Id: Ib3a33a7609530ef8c9f3f58fc607a61e8655c4b5

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
2024-02-27 23:03:49 -05:00
Andrew Dunham
23e9447871 tsweb: expose function to generate request IDs
For use in corp.

Updates tailscale/corp#2549

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I71debae1ce9ae48cf69cc44c2ab5c443fc3b2005
2024-02-27 18:57:53 -05:00
Mario Minardi
7912d76da0 client/web: update to typescript 5.3.3 (#11267)
Update typescript to 5.3.3. This is a major bump from the previous
version of 4.8.3. This also requires adding newer versions of
@typescript-eslint/eslint-plugin and @typescript-eslint/parser to our
resolutions as eslint-config-react-app pulls in versions that otherwise
do not support typescript 5.x.

eslint-config-react-app has not been updated in 2 years and is seemingly
abandoned, so we may wish to fork it or move to a different eslint config
in the future.

Updates https://github.com/tailscale/corp/issues/17810

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-02-27 14:17:30 -07:00
Andrew Dunham
c5abbcd4b4 wgengine/netstack: add a per-client limit for in-flight TCP forwards
This is a fun one. Right now, when a client is connecting through a
subnet router, here's roughly what happens:

1. The client initiates a connection to an IP address behind a subnet
   router, and sends a TCP SYN
2. The subnet router gets the SYN packet from netstack, and after
   running through acceptTCP, starts DialContext-ing the destination IP,
   without accepting the connection¹
3. The client retransmits the SYN packet a few times while the dial is
   in progress, until either...
4. The subnet router successfully establishes a connection to the
   destination IP and sends the SYN-ACK back to the client, or...
5. The subnet router times out and sends a RST to the client.
6. If the connection was successful, the client ACKs the SYN-ACK it
   received, and traffic starts flowing

As a result, the notification code in forwardTCP never notices when a
new connection attempt is aborted, and it will wait until either the
connection is established, or until the OS-level connection timeout is
reached and it aborts.

To mitigate this, add a per-client limit on how many in-flight TCP
forwarding connections can be in-progress; after this, clients will see
a similar behaviour to the global limit, where new connection attempts
are aborted instead of waiting. This prevents a single misbehaving
client from blocking all other clients of a subnet router by ensuring
that it doesn't starve the global limiter.

Also, bump the global limit again to a higher value.

¹ We can't accept the connection before establishing a connection to the
remote server since otherwise we'd be opening the connection and then
immediately closing it, which breaks a bunch of stuff; see #5503 for
more details.

Updates tailscale/corp#12184

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I76e7008ddd497303d75d473f534e32309c8a5144
2024-02-27 15:25:40 -05:00
Claire Wang
352c1ac96c tailcfg: add latitude, longitude for node location (#11162)
Updates tailscale/corp#17590

Signed-off-by: Claire Wang <claire@tailscale.com>
2024-02-27 15:02:06 -05:00
Irbe Krumina
95dcc1745b cmd/k8s-operator: reconcile tailscale Ingresses when their backend Services change. (#11255)
This is so that if a backend Service gets created after the Ingress, it gets picked up by the operator.

Updates tailscale/tailscale#11251

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
Co-authored-by: Anton Tolchanov <1687799+knyar@users.noreply.github.com>
2024-02-27 15:19:53 +00:00
Irbe Krumina
303125d96d cmd/k8s-operator: configure all proxies with declarative config (#11238)
Containerboot container created for operator's ingress and egress proxies
are now always configured by passing a configfile to tailscaled
(tailscaled --config <configfile-path>.
It does not run 'tailscale set' or 'tailscale up'.
Upgrading existing setups to this version as well as
downgrading existing setups at this version works.

Updates tailscale/tailscale#10869

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-02-27 15:14:09 +00:00
Irbe Krumina
45d27fafd6 cmd/k8s-operator,k8s-operator,go.{mod,sum},tstest/tools: add Tailscale Kubernetes operator API docs (#11246)
Add logic to autogenerate CRD docs.
.github/workflows/kubemanifests.yaml CI workflow will fail if the doc is out of date with regard to the current CRDs.
Docs can be refreshed by running make kube-generate-all.

Updates tailscale/tailscale#11023

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-02-27 14:51:53 +00:00
Percy Wegmann
05acf76392 tailfs: fix race condition in tailfs_test
Ues a noop authenticator to avoid potential races in gowebdav's
built-in authenticator.

Fixes #11259

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-27 08:31:04 -06:00
Keli
086ef19439 scripts/installer.sh: auto-start tailscale on Alpine (#11214)
On Alpine, we add the tailscale service but fail to call start.
This means that tailscale does not start up until the user reboots the machine.

Fixes #11161

Signed-off-by: Keli Velazquez <keli@tailscale.com>
2024-02-27 09:17:12 -05:00
Brad Fitzpatrick
1cf85822d0 ipn/ipnstate, wgengine/wgint: add handshake attempts accessors
Not yet used. This is being made available so magicsock/wgengine can
use it to ignore certain sends (UDP + DERP) later on at least mobile,
letting wireguard-go think it's doing its full attempt schedule, but
we can cut it short conditionally based on what we know from the
control plane.

Updates #7617

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Change-Id: Ia367cf6bd87b2aeedd3c6f4989528acdb6773ca7
2024-02-26 19:09:12 -08:00
Brad Fitzpatrick
eb28818403 wgengine: make pendOpen time later, after dup check
Otherwise on OS retransmits, we'd make redundant timers in Go's timer
heap that upon firing just do nothing (well, grab a mutex and check a
map and see that there's nothing to do).

Updates #cleanup

Change-Id: Id30b8b2d629cf9c7f8133a3f7eca5dc79e81facb
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-26 19:09:12 -08:00
Brad Fitzpatrick
219efebad4 wgengine: reduce critical section
No need to hold wgLock while using the device to LookupPeer;
that has its own mutex already.

Updates #cleanup

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Change-Id: Ib56049fcc7163cf5a2c2e7e12916f07b4f9d67cb
2024-02-26 19:09:12 -08:00
Brad Fitzpatrick
9a8c2f47f2 types/key: remove copy returning array by value
It's unnecessary. Returning an array value is already a copy.

Updates #cleanup

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Change-Id: If7f350b61003ea08f16a531b7b4e8ae483617939
2024-02-26 19:09:12 -08:00
Anton Tolchanov
8cc5c51888 health: warn about reverse path filtering and exit nodes
When reverse path filtering is in strict mode on Linux, using an exit
node blocks all network connectivity. This change adds a warning about
this to `tailscale status` and the logs.

Example in `tailscale status`:

```
- not connected to home DERP region 22
- The following issues on your machine will likely make usage of exit nodes impossible: [interface "eth0" has strict reverse-path filtering enabled], please set rp_filter=2 instead of rp_filter=1; see https://github.com/tailscale/tailscale/issues/3310
```

Example in the logs:
```
2024/02/21 21:17:07 health("overall"): error: multiple errors:
	not in map poll
	The following issues on your machine will likely make usage of exit nodes impossible: [interface "eth0" has strict reverse-path filtering enabled], please set rp_filter=2 instead of rp_filter=1; see https://github.com/tailscale/tailscale/issues/3310
```

Updates #3310

Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-02-27 00:43:01 +00:00
Nick Khyl
7ef1fb113d cmd/tailscaled, ipn/ipnlocal, wgengine: shutdown tailscaled if wgdevice is closed
Tailscaled becomes inoperative if the Tailscale Tunnel wintun adapter is abruptly removed.
wireguard-go closes the device in case of a read error, but tailscaled keeps running.
This adds detection of a closed WireGuard device, triggering a graceful shutdown of tailscaled.
It is then restarted by the tailscaled watchdog service process.

Fixes #11222

Signed-off-by: Nick Khyl <nickk@tailscale.com>
2024-02-26 14:45:35 -06:00
Nick Khyl
b42b9817b0 net/dns: do not wait for the interface registry key to appear if the windowsManager is being closed
The WinTun adapter may have been removed by the time we're closing
the dns.windowsManager, and its associated interface registry key might
also have been deleted. We shouldn't use winutil.OpenKeyWait and wait
for the interface key to appear when performing a cleanup as a part of
the windowsManager shutdown.

Updates #11222

Signed-off-by: Nick Khyl <nickk@tailscale.com>
2024-02-26 14:45:35 -06:00
OSS Updater
82c569a83a go.mod: update web-client-prebuilt module
Signed-off-by: OSS Updater <noreply+oss-updater@tailscale.com>
2024-02-26 13:26:47 -05:00
Sonia Appasamy
95f26565db client/web: use grants on web UI frontend
Starts using peer capabilities to restrict the management client
on a per-view basis. This change also includes a bulky cleanup
of the login-toggle.tsx file, which was getting pretty unwieldy
in its previous form.

Updates tailscale/corp#16695

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-02-26 12:59:37 -05:00
Sonia Appasamy
9aa704a05d client/web: restrict serveAPI endpoints to peer capabilities
This change adds a new apiHandler struct for use from serveAPI
to aid with restricting endpoints to specific peer capabilities.

Updates tailscale/corp#16695

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-02-26 12:59:37 -05:00
Anton Tolchanov
cd9cf93de6 wgengine/netstack: expose TCP forwarder drops via clientmetrics
- add a clientmetric with a counter of TCP forwarder drops due to the
  max attempts;
- fix varz metric types, as they are all counters.

Updates #8210

Signed-off-by: Anton Tolchanov <anton@tailscale.com>
2024-02-26 17:32:34 +00:00
Percy Wegmann
50fb8b9123 tailfs: replace webdavfs with reverse proxies
Instead of modeling remote WebDAV servers as actual
webdav.FS instances, we now just proxy traffic to them.
This not only simplifies the code, but it also allows
WebDAV locking to work correctly by making sure locks are
handled by the servers that need to (i.e. the ones actually
serving the files).

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-26 09:30:22 -06:00
Brad Fitzpatrick
e1bd7488d0 all: remove LenIter, use Go 1.22 range-over-int instead
Updates #11058
Updates golang/go#65685

Change-Id: Ibb216b346e511d486271ab3d84e4546c521e4e22
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-25 12:29:45 -08:00
mrrfv
ff1391a97e net/dns/publicdns: add Mullvad family DNS to the list of known DoH servers
Adds the new Mullvad family DNS server to the known DNS over HTTPS server list.

Signed-off-by: mrrfv <rm-rfv-no-preserve-root@protonmail.com>
2024-02-25 07:51:33 -08:00
Brad Fitzpatrick
6ad6d6b252 wgengine/wglog: add TS_DEBUG_RAW_WGLOG envknob for raw wg logs
Updates #7617 (part of debugging it)

Change-Id: I1bcbdcf0f929e3bcf83f244b1033fd438aa6dac1
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-24 14:59:48 -08:00
Brad Fitzpatrick
8b9474b06a wgengine/wgcfg: don't send UAPI to disable keep-alives on new peers
That's already the default. Avoid the overhead of writing it on one
side and reading it on the other to do nothing.

Updates #cleanup (noticed while researching something else)

Change-Id: I449c88a022271afb9be5da876bfaf438fe5d3f58
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-24 14:22:56 -08:00
James Tucker
8d0d46462b net/dns: timeout DOH requests after 10s without response headers
If a client socket is remotely lost but the client is not sent an RST in
response to the next request, the socket might sit in RTO for extended
lengths of time, resulting in "no internet" for users. Instead, timeout
after 10s, which will close the underlying socket, recovering from the
situation more promptly.

Updates #10967

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-23 23:08:12 -08:00
James Tucker
0c5e65eb3f cmd/derper: apply TCP keepalive and timeout to TLS as well
I missed a case in the earlier patch, and so we're still sending 15s TCP
keepalive for TLS connections, now adjusted there too.

Updates tailscale/corp#17587
Updates #3363

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-23 19:19:38 -08:00
James Tucker
c9b6d19fc9 .github/workflows: fix typo in XDG_CACHE_HOME
This appears to be one of the contributors to this CI target regularly
entering a bad state with a partially written toolchain.

Updates #self

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-23 16:44:29 -08:00
James Tucker
651c4899ac net/interfaces: reduce & cleanup logs on iOS
We don't need a log line every time defaultRoute is read in the good
case, and we now only log default interface updates that are actually
changes.

Updates #3363

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-23 16:37:06 -08:00
Andrea Gottardo
c8c999d7a9 cli/debug: rename DERP debug mode (#11220)
Renames a debug flag in the CLI.

Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
2024-02-23 15:48:37 -08:00
Mario Minardi
ac281dd493 client/web: update vite and vitest to latest versions (#11200)
Update vite to 5.1.4, and vitest to 1.3.1 (their latest versions). Also
remove vite-plugin-rewrite-all as this is no longer necessary with vite
5.x and has a dependency on vite 4.x.

Updates https://github.com/tailscale/corp/issues/17715

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-02-23 14:50:41 -07:00
Percy Wegmann
15b2c674bf cmd/tailscale: add node attribute instructions to share command help
This adds details on how to configure node attributes to allow
sharing and accessing shares.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-23 13:13:01 -06:00
James Tucker
131f9094fd wgengine/wglog: quieten WireGuard logs for allowedips
An increasing number of users have very large subnet route
configurations, which can produce very large amounts of log data when
WireGuard is reconfigured. The logs don't contain the actual routes, so
they're largely useless for diagnostics, so we'll just suppress them.

Fixes tailscale/corp#17532

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-23 10:06:22 -08:00
Andrew Dunham
e8d2fc7f7f net/tshttpproxy: log when we're using a proxy
Updates #11196

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Id6334c10f52f4cfbda9f03dc8096ab7a6c54a088
2024-02-22 19:22:50 -05:00
Mario Minardi
713d2928b1 client/web: update plugin-react-swc to latest version (#11199)
Update plugin-react-swc to the latest version (3.6.0) ahead of updating vite to 5.x.

Updates https://github.com/tailscale/corp/issues/17715

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-02-22 14:00:36 -07:00
Mario Minardi
72140da000 client/web: update vite-plugin-svgr to latest version (#11197)
Update vite-plugin-svgr to the latest version (4.2.0) ahead of updating
vite to 5.x. This is a major version bump from our previous 3.x, and
requires changing the import paths used for SVGs.

Updates https://github.com/tailscale/corp/issues/17715

Signed-off-by: Mario Minardi <mario@tailscale.com>
2024-02-22 13:16:44 -07:00
James Tucker
edbad6d274 cmd/derper: add user timeout and reduce TCP keepalive
The derper sends an in-protocol keepalive every 60-65s, so frequent TCP
keepalives are unnecessary. In this tuning TCP keepalives should never
occur for a DERP client connection, as they will send an L7 keepalive
often enough to always reset the TCP keepalive timer. If however a
connection does not receive an ACK promptly it will now be shutdown,
which happens sooner than it would with a normal TCP keepalive tuning.

This re-tuning reduces the frequency of network traffic from derp to
client, reducing battery cost.

Updates tailscale/corp#17587
Updates #3363

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-22 11:22:08 -08:00
Andrea Gottardo
0359c2f94e util/syspolicy: add 'ResetToDefaults' (#11194)
Updates ENG-2133. Adds the ResetToDefaults visibility policy currently only available on macOS, so that the Windows client can read its value.

Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
2024-02-22 10:10:31 -08:00
Brad Fitzpatrick
10d130b845 cmd/derper, derp, tailcfg: add admission controller URL option
So derpers can check an external URL for whether to permit access
to a certain public key.

Updates tailscale/corp#17693

Change-Id: I8594de58f54a08be3e2dbef8bcd1ff9b728ab297
Co-authored-by: Maisem Ali <maisem@tailscale.com>
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-21 16:57:45 -08:00
Brad Fitzpatrick
2988c1ec52 derp: plumb context to Server.verifyClient
Updates tailscale/corp#17693

Change-Id: If17e02c77d5ad86b820e639176da2d3e61296bae
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-21 16:24:44 -08:00
Paul Scott
7708ab68c0 cmd/tailscale/cli: pass "-o 'CanonicalizeHostname no'" to ssh
Fixes #10348

Signed-off-by: Paul Scott <paul@tailscale.com>
2024-02-21 22:34:34 +00:00
Percy Wegmann
91a1019ee2 cmd/testwrapper: apply results of all unit tests to coverage for all packages
This allows coverage from tests that hit multiple packages at once
to be reflected in all those packages' coverage.

Updates #cleanup

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-21 13:08:17 -06:00
Andrea Gottardo
d756622432 util/syspolicy: add ManagedBy keys for Windows (#11183) 2024-02-20 15:08:06 -08:00
James Tucker
8fe504241d net/ktimeout: add a package to set TCP user timeout
Setting a user timeout will be a more practical tuning knob for a number
of endpoints, this provides a way to set it.

Updates tailscale/corp#17587

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-20 10:49:58 -08:00
Brad Fitzpatrick
a4a909a20b prober: add TLS probe constructor to split dial addr from cert name
So we can probe load balancers by their unique DNS name but without
asking for that cert name.

Updates tailscale/corp#13050

Change-Id: Ie4c0a2f951328df64281ed1602b4e624e3c8cf2e
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-19 09:03:13 -08:00
Brad Fitzpatrick
794af40f68 ipn/ipnlocal: remove ancient transition mechanism for https certs
And confusing error message that duplicated the valid cert domains.

Fixes tailscale/corp#15876

Change-Id: I098bc45d83c8d1e0a233dcdf3188869cce66e128
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-17 10:33:11 -08:00
James Tucker
6c3899e6ee logpolicy: allow longer idle log upload connections
From a packet trace we have seen log connections being closed
prematurely by the client, resulting in unnecessary extra TLS setup
traffic.

Updates #3363
Updates tailscale/corp#9230
Updates tailscale/corp#8564

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-16 18:06:09 -08:00
Andrew Dunham
70b7201744 net/dns: fix infinite loop when run on Amazon Linux 2023
This fixes an infinite loop caused by the configuration of
systemd-resolved on Amazon Linux 2023 and how that interacts with
Tailscale's "direct" mode. We now drop the Tailscale service IP from the
OS's "base configuration" when we detect this configuration.

Updates #7816

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I73a4ea8e65571eb368c7e179f36af2c049a588ee
2024-02-16 18:07:32 -05:00
Andrea Gottardo
44e337cc0e tool/gocross: pass flags for visionOS and visionOS Simulator (#11127)
Adds logic in gocross to detect environment variables and pass the right flags so that the backend can be built with the visionOS SDK.

Signed-off-by: Andrea Gottardo <andrea@tailscale.com>
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
2024-02-16 11:14:17 -08:00
Will Norris
6b582cb8b6 cmd/tailscale: support clickable IPv6 web client addresses
Instead of constructing the `ip:port` string ourselves, use
netip.AddrPortFrom which handles IPv6 correctly.

Updates #11164

Signed-off-by: Will Norris <will@tailscale.com>
2024-02-16 11:00:37 -08:00
Will Norris
24487815e1 cmd/tailscale: make web client URL clickable
Updates #11151

Signed-off-by: Will Norris <will@tailscale.com>
2024-02-16 10:28:34 -08:00
San
69f5664075 ipn/ipnlocal: fix doctor API endpoint (#11155)
Small fix to make sure doctor API endpoint returns correctly - I spotted it when checking my tailscaled node and noticed it was handled slightly different compare to the rest

Signed-off-by: San <santrancisco@users.noreply.github.com>
2024-02-16 12:17:34 -05:00
Percy Wegmann
3aca29e00e VERSION.txt: this is v1.61.0
Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-15 17:24:49 -06:00
Jason Barnett
4d668416b8 wgengine/router: fix ip rule restoration
Fixes #10857

Signed-off-by: Jason Barnett <J@sonBarnett.com>
2024-02-15 11:36:40 -05:00
Andrew Dunham
52f16b5d10 doctor/ethtool, ipn/ipnlocal: add ethtool bugreport check
Updates #11137

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Idbe862d80e428adb044249c47d9096b87f29d5d8
2024-02-15 10:17:05 -05:00
Patrick O'Doherty
38bba2d23a clientupdate: disable auto update on NixOS (#11136)
Updates #cleanup

NixOS packages are immutable and attempts to update via our tarball
mechanism will always fail as a result. Instead we now direct users to
update their nix channel or nixpkgs flake input to receive the latest
Tailscale release.

Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
2024-02-14 11:58:29 -08:00
Andrew Dunham
b7104cde4a util/topk: add package containing a probabilistic top-K tracker
This package uses a count-min sketch and a heap to track the top K items
in a stream of data. Tracking a new item and adding a count to an
existing item both require no memory allocations and is at worst
O(log(k)) complexity.

Change-Id: I0553381be3fef2470897e2bd806d43396f2dbb36
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
2024-02-14 13:28:58 -05:00
Flakes Updater
7ad2bb87a6 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-02-13 19:39:53 -08:00
Brad Fitzpatrick
61a1644c2a go.mod, all: move away from inet.af domain seized by Taliban
Updates inetaf/tcpproxy#39

Change-Id: I7fee276b116bd08397347c6c949011d76a2842cf
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-13 19:21:09 -08:00
Andrew Dunham
b0e96a6c39 net/dns: log more info when openresolv commands fail
Updates #11129

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ic594868ba3bc31f6d3b0721ecba4090749a81f7f
2024-02-13 20:48:54 -05:00
Nathan Woodburn
7c0651aea6 scripts/installer.sh: add tuxedoOS to the Ubuntu copies
Signed-off-by: Nathan Woodburn <github@nathan.woodburn.au>
2024-02-13 15:37:15 -08:00
Patrick O'Doherty
256ecd0e8f Revert "tsweb: update ServeMux matching to 1.22.0 syntax (#11090)" (#11125)
This reverts commit 30c9189ed3.

Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
2024-02-13 10:49:36 -08:00
Aaron Klotz
f7acbefbbb wgengine/router: make the Windows ifconfig implementation reuse existing MibIPforwardRow2 when possible
Looking at profiles, we spend a lot of time in winipcfg.LUID.DeleteRoute
looking up the routing table entry for the provided RouteData.

But we already have the row! We previously obtained that data via the full
table dump we did in getInterfaceRoutes. We can make this a lot faster by
hanging onto a reference to the wipipcfg.MibIPforwardRow2 and executing
the delete operation directly on that.

Fixes #11123

Signed-off-by: Aaron Klotz <aaron@tailscale.com>
2024-02-13 11:17:01 -07:00
Patrick O'Doherty
30c9189ed3 tsweb: update ServeMux matching to 1.22.0 syntax (#11090)
* tsweb: update ServeMux matching to 1.22.0 syntax

Updates #cleanup

Go 1.22.0 introduced the ability to use more expressive routing patterns
that include HTTP method when constructing ServeMux entries.
Applications that attempted to use these patterns in combination with
the old `tsweb.Debugger` would experience a panic as Go would not permit
the use of matching rules with mixed level of specificity. We now
specify the method for each `/debug` handler to prevent
incompatibilities.

Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
2024-02-13 09:56:00 -08:00
Irbe Krumina
5bd19fd3e3 cmd/k8s-operator,k8s-operator: proxy configuration mechanism via a new ProxyClass custom resource (#11074)
* cmd/k8s-operator,k8s-operator: introduce proxy configuration mechanism via ProxyClass custom resource.

ProxyClass custom resource can be used to specify customizations
for the proxy resources created by the operator.

Add a reconciler that validates ProxyClass resources
and sets a Ready condition to True or False with a corresponding reason and message.
This is required because some fields (labels and annotations)
require complex validations that cannot be performed at custom resource apply time.
Reconcilers that use the ProxyClass to configure proxy resources are expected to
verify that the ProxyClass is Ready and not proceed with resource creation
if configuration from a ProxyClass that is not yet Ready is required.

If a tailscale ingress/egress Service is annotated with a tailscale.com/proxy-class annotation, look up the corresponding ProxyClass and, if it is Ready, apply the configuration from the ProxyClass to the proxy's StatefulSet.

If a tailscale Ingress has a tailscale.com/proxy-class annotation
and the referenced ProxyClass custom resource is available and Ready,
apply configuration from the ProxyClass to the proxy resources
that will be created for the Ingress.

Add a new .proxyClass field to the Connector spec.
If connector.spec.proxyClass is set to a ProxyClass that is available and Ready,
apply configuration from the ProxyClass to the proxy resources created for the Connector.

Ensure that when Helm chart is packaged, the ProxyClass yaml is added to chart templates. Ensure that static manifest generator adds ProxyClass yaml to operator.yaml. Regenerate operator.yaml


Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-02-13 05:27:54 +00:00
Brad Fitzpatrick
f7f496025a types/views: add test that LenIter doesn't allocate
For a second we thought this was allocating but we were looking
at a CPU profile (which showed calls to mallocgc view makeslice)
instead of the alloc profile.

Updates golang/go#65685 (which if fixed wouldn't have confused us)

Change-Id: Ic0132310d52d8a65758a516142525339aa23b1ed
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-12 18:04:06 -08:00
Percy Wegmann
c42a4e407a tailfs: listen for local clients only on 100.100.100.100
FileSystemForLocal was listening on the node's Tailscale address,
which potentially exposes the user's view of TailFS shares to other
Tailnet users. Remote nodes should connect to exported shares via
the peerapi.

This removes that code so that FileSystemForLocal is only avaialable
on 100.100.100.100:8080.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-12 14:08:00 -06:00
Percy Wegmann
d0ef3a25df cmd/tailscale: hide share subcommand
Fixes #1115

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-12 14:03:01 -06:00
David Anderson
58b8f78e7e flake.nix: build tailscale with go 1.22
Updates #cleanup

Signed-off-by: David Anderson <danderson@tailscale.com>
2024-02-11 20:43:40 -08:00
Maisem Ali
370ecb4654 tailcfg: remove UserProfile.Groups
Removing as per go/group-all-the-things.

Updates tailscale/corp#17445

Signed-off-by: Maisem Ali <maisem@tailscale.com>
2024-02-11 09:44:11 -08:00
Andrew Dunham
c1c50cfcc0 util/cloudenv: add support for DigitalOcean
Updates #4984

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ib229eb40af36a80e6b0fd1dd0cabb07f0d50a7d1
2024-02-10 14:36:20 -05:00
Percy Wegmann
55b372a79f tailscaled: revert to using pointers for subcommands
As part of #10631, we stopped using function pointers for subcommands,
preventing us from registering platform-specific installSystemDaemon
and uninstallSystemDaemon subcommands.

Fixes #11099

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-10 11:55:24 -06:00
Percy Wegmann
87154a2f88 tailfs: fix startup issues on windows
Starts TailFS for Windows too, initializes shares on startup.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-09 22:13:27 -06:00
Percy Wegmann
ddcffaef7a tailfs: disable TailFSForLocal via policy
Adds support for node attribute tailfs:access. If this attribute is
not present, Tailscale will not accept connections to the local TailFS
server at 100.100.100.100:8080.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-09 20:00:42 -06:00
Percy Wegmann
abab0d4197 tailfs: clean up naming and package structure
- Restyles tailfs -> tailFS
- Defines interfaces for main TailFS types
- Moves implemenatation of TailFS into tailfsimpl package

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-09 20:00:42 -06:00
dependabot[bot]
79b547804b build(deps-dev): bump vite from 4.4.9 to 4.5.2 in /client/web
Bumps [vite](https://github.com/vitejs/vite/tree/HEAD/packages/vite) from 4.4.9 to 4.5.2.
- [Release notes](https://github.com/vitejs/vite/releases)
- [Changelog](https://github.com/vitejs/vite/blob/v4.5.2/packages/vite/CHANGELOG.md)
- [Commits](https://github.com/vitejs/vite/commits/v4.5.2/packages/vite)

---
updated-dependencies:
- dependency-name: vite
  dependency-type: direct:development
...

Signed-off-by: dependabot[bot] <support@github.com>
2024-02-09 18:42:56 -05:00
James Tucker
24bac27632 util/rands: add Shuffle and Perm functions with on-stack RNG state
The new math/rand/v2 package includes an m-local global random number
generator that can not be reseeded by the user, which is suitable for
most uses without the RNG pools we have in a number of areas of the code
base.

The new API still does not have an allocation-free way of performing a
seeded operations, due to the long term compiler bug around interface
parameter escapes, and the Source interface.

This change introduces the two APIs that math/rand/v2 can not yet
replace efficiently: seeded Perm() and Shuffle() operations. This
implementation chooses to use the PCG random source from math/rand/v2,
as with sufficient compiler optimization, this source should boil down
to only two on-stack registers for random state under ideal conditions.

Updates #17243

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-09 15:19:27 -08:00
Sonia Appasamy
2bb837a9cf client/web: only check policy caps for tagged nodes
For user-owned nodes, only the owner is ever allowed to manage the
node.

Updates tailscale/corp#16695

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-02-09 18:17:14 -05:00
Andrew Dunham
6f6383f69e .github: fuzzing is now unbroken
Updates #cleanup

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I10dca601c79411b412180a46b3f82136e40544b0
2024-02-09 17:02:58 -05:00
Keisuke Umegaki
7039c06d9b fix toolchain not available error (#11083)
Relates to golang/go#62278
Updates #11058

Signed-off-by: keisku <keisuke.umegaki.630@gmail.com>
2024-02-09 16:46:32 -05:00
Patrick O'Doherty
7c52b27daf Revert "tsweb: update ServeMux matching to 1.22.0 syntax (#11087)" (#11089)
This reverts commit 291f91d164.

Updates #cleanup

This PR needs additional changes to the registration of child handlers under /debug

Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
2024-02-09 12:23:49 -08:00
Patrick O'Doherty
291f91d164 tsweb: update ServeMux matching to 1.22.0 syntax (#11087)
Updates #cleanup

Go 1.22.0 introduced the ability to use more expressive routing patterns
that include HTTP method when constructing ServeMux entries.
Applications that attempted to use these patterns in combination with
the old `tsweb.Debugger` would experience a panic as Go would not permit
the use of matching rules with mixed level of specificity.

Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
2024-02-09 11:27:05 -08:00
Jenny Zhang
c446451bfa cmd/gitops-pusher: only use OAuth creds if non-empty string
`os.LookupEnv` may return true if the variable is present in
the environment but an empty string. We should only attempt
to set OAuth Config if thsoe values are non-empty.

Updates gitops-acl-action#33

Signed-off-by: Jenny Zhang <jz@tailscale.com>
2024-02-09 10:55:59 -05:00
Percy Wegmann
993acf4475 tailfs: initial implementation
Add a WebDAV-based folder sharing mechanism that is exposed to local clients at
100.100.100.100:8080 and to remote peers via a new peerapi endpoint at
/v0/tailfs.

Add the ability to manage folder sharing via the new 'share' CLI sub-command.

Updates tailscale/corp#16827

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-02-09 09:13:51 -06:00
Joe Tsai
2e404b769d all: use new AppendEncode methods available in Go 1.22 (#11079)
Updates #cleanup

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-02-08 17:55:03 -08:00
Joe Tsai
94a4f701c2 all: use reflect.TypeFor now available in Go 1.22 (#11078)
Updates #cleanup

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-02-08 17:34:22 -08:00
Joe Tsai
efddad7d7d util/deephash: cleanup TODO in TestHash (#11080)
Updates #cleanup

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-02-08 17:33:25 -08:00
Charlotte Brandhorst-Satzkorn
0f042b9814 cmd/tailscale/cli: fix exit node status output (#11076)
This change fixes the format of tailscale status output when location
based exit nodes are present.

Fixes #11065

Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
2024-02-08 14:29:01 -08:00
Andrea Gottardo
6c79f55d48 ipnlocal: force-regen new authURL when it is too old (#10971)
Fixes tailscale/support-escalations#23.

authURLs returned by control expire after 1 hour from creation. Customer reported that the Tailscale client on macOS would sending users to a stale authentication page when clicking on the `Login...` menu item. This can happen when clicking on Login after leaving the device unattended for several days. The device key expires, leading to the creation of a new authURL, however the client doesn't keep track of when the authURL was created. Meaning that `login-interactive` would send the user to an authURL that had expired server-side a long time before.

This PR ensures that whenever `login-interactive` is called via LocalAPI, an authURL that is too old won't be used. We force control to give us a new authURL whenever it's been more than 30 minutes since the last authURL was sent down from control.



Apply suggestions from code review




Set interval to 6 days and 23 hours

Signed-off-by: Andrea Gottardo <andrea@tailscale.com>
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
2024-02-08 13:04:01 -08:00
Sonia Appasamy
1217f655c0 cmd/dist: update logs for synology builds
Update logs for synology builds to more clearly callout which variant
is being built. The two existing variants are:

1. Sideloaded (can be manual installed on a device by anyone)
2. Package center distribution (by the tailscale team)

Updates #cleanup

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-02-08 14:36:55 -05:00
OSS Updater
664b861cd4 go.mod: update web-client-prebuilt module
Signed-off-by: OSS Updater <noreply+oss-updater@tailscale.com>
2024-02-08 10:57:10 -08:00
Will Norris
6f0c5e0c05 client/web: use smart quotes in web UI frontend
add the curly-quotes eslint plugin (same that we use for the admin
panel), and fix existing straight quotes in the current web UI.

Updates #cleanup

Signed-off-by: Will Norris <will@tailscale.com>
2024-02-08 10:13:46 -08:00
Will Norris
128c99d4ae client/web: add new readonly mode
The new read-only mode is only accessible when running `tailscale web`
by passing a new `-readonly` flag. This new mode is identical to the
existing login mode with two exceptions:

 - the management client in tailscaled is not started (though if it is
   already running, it is left alone)

 - the client does not prompt the user to login or switch to the
   management client. Instead, a message is shown instructing the user
   to use other means to manage the device.

Updates #10979

Signed-off-by: Will Norris <will@tailscale.com>
2024-02-08 10:11:23 -08:00
License Updater
9f0eaa4464 licenses: update win/apple licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-02-08 08:42:15 -08:00
License Updater
78f257d9f8 licenses: update android licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-02-08 08:41:12 -08:00
License Updater
5486d8aaf9 licenses: update tailscale{,d} licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-02-08 08:40:53 -08:00
Irbe Krumina
a6cc2fdc3e cmd/{containerboot,k8s-operator/deploy/manifests}: optionally allow proxying cluster traffic to a cluster target via ingress proxy (#11036)
* cmd/containerboot,cmd/k8s-operator/deploy/manifests: optionally forward cluster traffic via ingress proxy.

If a tailscale Ingress has tailscale.com/experimental-forward-cluster-traffic-via-ingress annotation, configure the associated ingress proxy to have its tailscale serve proxy to listen on Pod's IP address. This ensures that cluster traffic too can be forwarded via this proxy to the ingress backend(s).

In containerboot, if EXPERIMENTAL_PROXY_CLUSTER_TRAFFIC_VIA_INGRESS is set to true
and the node is Kubernetes operator ingress proxy configured via Ingress,
make sure that traffic from within the cluster can be proxied to the ingress target.

Updates tailscale/tailscale#10499

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-02-08 06:45:42 +00:00
Flakes Updater
2404b1444e go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-02-07 18:18:32 -08:00
Brad Fitzpatrick
c424e192c0 .github/workflows: temporarily disable broken oss-fuzz action
Updates #11064
Updates #11058

Change-Id: I63acc13dece3379a0b2df573afecfd245b7cd6c2
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-07 18:10:15 -08:00
Brad Fitzpatrick
2bd3c1474b util/cmpx: delete now that we're using Go 1.22
Updates #11058

Change-Id: I09dea8e86f03ec148b715efca339eab8b1f0f644
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-07 18:10:15 -08:00
Brad Fitzpatrick
5ea071186e Dockerfile: use Go 1.22
Updates #11058

Change-Id: I0f63be498be33d71bd90b7956f9fe9666fd7a696
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-07 18:10:15 -08:00
Brad Fitzpatrick
9612001cf4 .github/workflows: update golangci-lint for Go 1.22
Updates #11058

Change-Id: I3785c1f1bea4a4663e7e5fb6d209d3caedae436d
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-07 18:10:15 -08:00
Brad Fitzpatrick
b6153efb7d go.mod, README.md: use Go 1.22
Updates #11058

Change-Id: I95eecdc7afe2b5f8189016fdb8a773f78e9f5c42
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2024-02-07 18:10:15 -08:00
Tom DNetto
653721541c tsweb: normalize passkey identities in bucketed stats
Signed-off-by: Tom DNetto <tom@tailscale.com>
Updates: corp#17075
2024-02-07 16:02:36 -08:00
Tom DNetto
8d6d9d28ba tsweb: normalize common StableID's in bucketed stats, export as LabelMap
Signed-off-by: Tom DNetto <tom@tailscale.com>
Updates: corp#17075
2024-02-07 15:36:39 -08:00
James Tucker
e0762fe331 words: add a list of things you should yahoo!
Updates #self

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-07 14:47:20 -08:00
James Tucker
0b16620b80 .github/workflows: add privileged tests workflow
We had missed regressions from privileged tests not running, now they
can run.

Updates #cleanup
Signed-off-by: James Tucker <james@tailscale.com>
2024-02-07 14:45:22 -08:00
James Tucker
0f5e031133 appc: optimize dns response observation for large route tables
Advertise DNS discovered addresses as a single preference update rather
than one at a time.

Sort the list of observed addresses and use binary search to consult the
list.

Updates tailscale/corp#16636

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-07 14:11:41 -08:00
Andrew Lytvynov
db3776d5bf go.toolchain.rev: bump to Go 1.22.0 (#11055)
Updates #cleanup

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-02-07 14:57:57 -07:00
Tom DNetto
af931dcccd tsweb: replace domains/emails in paths when bucketing stats
Signed-off-by: Tom DNetto <tom@tailscale.com>
Updates: corp#17075
2024-02-07 13:31:59 -08:00
Tom DNetto
36efc50817 tsweb: implementing bucketed statistics for started/finished counts
Signed-off-by: Tom DNetto <tom@tailscale.com>
Updates: corp#17075
2024-02-06 16:54:17 -08:00
Maisem Ali
b752bde280 types/views: add SliceMapKey[T]
views.Slice are meant to be immutable, and if used as such it
is at times desirable to use them as a key in a map. For non-viewed
slices it was kinda doable by creating a custom key struct but views.Slice
didn't allow for the same so add a method to create that struct here.

Updates tailscale/corp#17122

Signed-off-by: Maisem Ali <maisem@tailscale.com>
2024-02-06 12:50:28 -08:00
kari-ts
5595b61b96 ipn/localapi: more http status cleanup (#10995)
Use Http.StatusOk instead of 200

Updates #cleanup
2024-02-05 09:34:41 -08:00
Chris Palmer
a633a30711 cmd/hello: link to the Hello KB article (#11022)
Fixes https://github.com/tailscale/corp/issues/17104

Signed-off-by: Chris Palmer <cpalmer@tailscale.com>
2024-02-02 15:48:31 -08:00
Joe Tsai
60657ac83f util/deephash: tighten up SelfHasher API (#11012)
Providing a hash.Block512 is an implementation detail of how deephash
works today, but providing an opaque type with mostly equivalent API
(i.e., HashUint8, HashBytes, etc. methods) is still sensible.
Thus, define a public Hasher type that exposes exactly the API
that an implementation of SelfHasher would want to call.
This gives us freedom to change the hashing algorithm of deephash
at some point in the future.

Also, this type is likely going to be called by types that are
going to memoize their own hash results, we additionally add
a HashSum method to simplify this use case.

Add documentation to SelfHasher on how a type might implement it.

Updates: corp#16409

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-02-01 17:07:41 -08:00
Joe Tsai
84f8311bcd util/deephash: document pathological deephash behavior (#11010)
Updates #cleanup

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-02-01 13:49:36 -08:00
James Tucker
ba70cbb930 ipn/ipnlocal: fix app connector route advertisements on exit nodes
If an app connector is also configured as an exit node, it should still
advertise discovered routes that are not covered by advertised routes,
excluding the exit node routes.

Updates tailscale/corp#16928

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-01 11:56:24 -08:00
James Tucker
e1a4b89dbe appc,ipn/ipnlocal: add app connector routes if any part of a CNAME chain is routed
If any domain along a CNAME chain matches any of the routed domains, add
routes for the discovered domains.

Fixes tailscale/corp#16928

Signed-off-by: James Tucker <james@tailscale.com>
2024-02-01 11:43:07 -08:00
Tom DNetto
2aeef4e610 util/deephash: implement SelfHasher to allow types to hash themselves
Updates: corp#16409
Signed-off-by: Tom DNetto <tom@tailscale.com>
2024-02-01 10:18:03 -08:00
James Tucker
b4b2ec7801 ipn/ipnlocal: fix pretty printing of multi-record peer DNS results
The API on the DNS record parser is slightly subtle and requires
explicit handling of unhandled records. Failure to advance previously
resulted in an infinite loop in the pretty responder for any reply that
contains a record other than A/AAAA/TXT.

Updates tailscale/corp#16928

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-31 15:59:17 -08:00
Percy Wegmann
fad6bae764 ipnlocal: log failure to get ssh host keys
When reporting ssh host keys to control, log a warning
if we're unable to get the SSH host keys.

Updates tailscale/escalations#21

Signed-off-by: Percy Wegmann <percy@tailscale.com>
2024-01-30 16:57:16 -06:00
Chris Palmer
9744ad47e3 cmd/hello: avoid deprecated apis (#10957)
Updates #cleanup

Signed-off-by: Chris Palmer <cpalmer@tailscale.com>
2024-01-29 14:18:49 -08:00
Will Norris
13f8a669d5 cmd/gitops-pusher: fix logic for checking credentials
gitops-pusher supports authenticating with an API key or OAuth
credentials (added in #7393). You shouldn't ever use both of those
together, so we error if both are set.

In tailscale/gitops-acl-action#24, OAuth support is being added to the
GitHub action. In that environment, both the TS_API_KEY and OAuth
variables will be set, even if they are empty values.  This causes an
error in gitops-pusher which expects only one to be set.

Update gitops-pusher to check that only one set of environment variables
are non-empty, rather than just checking if they are set.

Updates #7393

Signed-off-by: Will Norris <will@tailscale.com>
2024-01-29 13:01:29 -08:00
Charlotte Brandhorst-Satzkorn
cce189bde1 words: i like the direction this list is taking
Updates tailscale/corp#14698

Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
2024-01-26 19:42:37 -08:00
Andrew Lytvynov
fbfc3b7e51 cmd/tailscale/cli: run Watch with NotifyNoPrivateKeys (#10950)
When running as non-root non-operator user, you get this error:
```
$ tailscale serve 8080
Access denied: watch IPN bus access denied, must set ipn.NotifyNoPrivateKeys when not running as admin/root or operator

Use 'sudo tailscale serve 8080' or 'tailscale up --operator=$USER' to not require root.
```

It should fail, but the error message is confusing.

With this fix:
```
$ tailscale serve 8080
sending serve config: Access denied: serve config denied

Use 'sudo tailscale serve 8080' or 'tailscale up --operator=$USER' to not require root.
```

Updates #cleanup

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-01-25 14:59:34 -07:00
James Tucker
0f3b2e7b86 util/expvarx: add a time and concurrency limiting expvar.Func wrapper
expvarx.SafeFunc wraps an expvar.Func with a time limit. On reaching the
time limit, calls to Value return nil, and no new concurrent calls to
the underlying expvar.Func will be started until the call completes.

Updates tailscale/corp#16999
Signed-off-by: James Tucker <james@tailscale.com>
2024-01-24 17:47:16 -08:00
Andrew Dunham
fd94d96e2b net/portmapper: support legacy "urn:dslforum-org" portmapping services
These are functionally the same as the "urn:schemas-upnp-org" services
with a few minor changes, and are still used by older devices. Support
them to improve our ability to obtain an external IP on such networks.

Updates #10911

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I05501fad9d6f0a3b8cf19fc95eee80e7d16cc2cf
2024-01-23 21:29:29 -05:00
Irbe Krumina
75f1d3e7d7 ipn/ipnlocal: fix failing test (#10937)
Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-23 19:02:07 +00:00
Irbe Krumina
6ee956333f ipn/ipnlocal: fix proxy path that matches mount point (#10864)
Don't append a trailing slash to a request path
to the reverse proxy that matches the mount point exactly.

Updates tailscale/tailscale#10730

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-23 18:12:56 +00:00
Jordan Whited
8b47322acc wgengine/magicsock: implement probing of UDP path lifetime (#10844)
This commit implements probing of UDP path lifetime on the tail end of
an active direct connection. Probing configuration has two parts -
Cliffs, which are various timeout cliffs of interest, and
CycleCanStartEvery, which limits how often a probing cycle can start,
per-endpoint. Initially a statically defined default configuration will
be used. The default configuration has cliffs of 10s, 30s, and 60s,
with a CycleCanStartEvery of 24h. Probing results are communicated via
clientmetric counters. Probing is off by default, and can be enabled
via control knob. Probing is purely informational and does not yet
drive any magicsock behaviors.

Updates #540

Signed-off-by: Jordan Whited <jordan@tailscale.com>
2024-01-23 09:37:32 -08:00
James Tucker
0e2cb76abe appc: add test to ensure that individual IPs are not removed during route updates
If control advised the connector to advertise a route that had already
been discovered by DNS it would be incorrectly removed. Now those routes
are preserved.

Updates tailscale/corp#16833

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-22 17:50:55 -08:00
Charlotte Brandhorst-Satzkorn
ce4553b988 appc,ipn/ipnlocal: optimize preference adjustments when routes update
This change allows us to perform batch modification for new route
advertisements and route removals. Additionally, we now handle the case
where newly added routes are covered by existing ranges.

This change also introduces a new appctest package that contains some
shared functions used for testing.

Updates tailscale/corp#16833

Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
2024-01-22 17:37:16 -08:00
Irbe Krumina
370ec6b46b cmd/k8s-operator: don't proceed with Ingress that has no valid backends (#10919)
Do not provision resources for a tailscale Ingress that has no valid backends.

Updates tailscale/tailscale#10910

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-22 19:20:23 +00:00
Andrew Dunham
b45089ad85 net/portmapper: handle cases where we have no supported clients
This no longer results in a nil pointer exception when we get a valid
UPnP response with no supported clients.

Updates #10911

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I6e3715a49a193ff5261013871ad7fff197a4d77e
2024-01-22 12:46:19 -05:00
James Tucker
4e822c031f go.toolchain.rev: bump Tailscale Go version to 1.21.6
Updates tailscale/go#83

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-19 18:30:35 -08:00
Flakes Updater
b787c27c00 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-01-19 18:24:58 -08:00
James Tucker
7e3bcd297e go.mod,wgengine/netstack: bump gvisor
Updates #8043

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-19 18:23:53 -08:00
David Anderson
17eae5b0d3 tool/gocross: force use of our custom toolchain
The new 'toolchain' directive in go.mod can sometimes force
the use of an upstream toolchain against our wishes. Concurrently,
some of our dependencies have added the 'toolchain' directive, which
transitively adds it to our own go.mod. Force all uses of gocross to
ignore that directive and stick to our customized toolchain.

Updates #cleanup

Signed-off-by: David Anderson <danderson@tailscale.com>
2024-01-19 18:23:21 -08:00
David Anderson
ae79b2e784 tsweb: add a helper to validate redirect URLs
We issue redirects in a few different places, it's time to have
a common helper to do target validation.

Updates tailscale/corp#16875

Signed-off-by: David Anderson <danderson@tailscale.com>
2024-01-19 17:57:54 -08:00
Claire Wang
213d696db0 magicsock: mute noisy expected peer mtu related error (#10870) 2024-01-19 20:04:22 -05:00
kari-ts
62b056d677 VERSION.txt: this is v1.59.0 (#10884)
* VERSION.txt: this is v1.58.0

Signed-off-by: kari-ts <kari@tailscale.com>

* VERSION.txt: this is v1.59.0

---------

Signed-off-by: kari-ts <kari@tailscale.com>
2024-01-19 17:03:50 -08:00
Flakes Updater
5b4eb47300 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-01-19 16:53:05 -08:00
James Tucker
457102d070 go.mod: bump most deps for start of cycle
Plan9 CI is disabled. 3p dependencies do not build for the target.
Contributor enthusiasm appears to have ceased again, and no usage has
been made.

Skipped gvisor, nfpm, and k8s.

Updates #5794
Updates #8043

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-19 16:51:39 -08:00
Andrew Dunham
7a0392a8a3 wgengine/netstack: expose gVisor metrics through expvar
When tailscaled is run with "-debug 127.0.0.1:12345", these metrics are
available at:
    http://localhost:12345/debug/metrics

Updates #8210

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I19db6c445ac1f8344df2bc1066a3d9c9030606f8
2024-01-19 18:36:54 -05:00
as2643
832e5c781d util/nocasemaps: add AppendSliceElem method to nocasemaps (#10871)
Updates #7667

Signed-off-by: Anishka Singh <anishkasingh66@gmail.com>
2024-01-19 15:30:12 -08:00
ChandonPierre
2ce596ea7a cmd/k8s-operator/deploy: allow modifying operator tags via Helm values
Updates tailscale/tailscale#10659

Signed-off-by: Chandon Pierre <cpierre@coreweave.com>
2024-01-19 21:22:23 +00:00
Andrew Dunham
2ac7c0161b util/slicesx: add Filter function
For use in corp, where we appear to have re-implemented this in a few
places with varying signatures.

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Id863a87e674f3caa87945519be8e09650e9c1d76
2024-01-19 11:17:05 -05:00
Irbe Krumina
2aec4f2c43 ./github/workflows/kubemanifests.yaml: fix the paths whose changes should trigger test runs (#10885)
Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-19 01:51:10 +00:00
James Tucker
8250582fe6 ipn/ipnlocal: make app connector configuration concurrent
If there are routes changes as a side effect of an app connector
configuration update, the connector configuration may want to reenter a
lock, so must be started asynchronously.

Updates tailscale/corp#16833
Signed-off-by: James Tucker <james@tailscale.com>
2024-01-18 12:26:58 -08:00
James Tucker
38a1cf748a control/controlclient,util/execqueue: extract execqueue into a package
This is a useful primitive for asynchronous execution of ordered work I
want to use in another change.

Updates tailscale/corp#16833
Signed-off-by: James Tucker <james@tailscale.com>
2024-01-18 12:08:13 -08:00
Flakes Updater
32f01acc79 go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-01-17 20:06:58 -08:00
James Tucker
24df1ef1ee appc,ipn/ipnlocal,types/appctype: implement control provided routes
Control can now send down a set of routes along with the domains, and
the routes will be advertised, with any newly overlapped routes being
removed to reduce the size of the routing table.

Fixes tailscale/corp#16833
Signed-off-by: James Tucker <james@tailscale.com>
2024-01-17 14:40:09 -08:00
Andrea Gottardo
543e7ed596 licenses: mention tvOS in apple.md (#10872)
Signed-off-by: Andrea Gottardo <andrea@tailscale.com>
2024-01-16 18:32:20 -08:00
License Updater
3eba895293 licenses: update tailscale{,d} licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-01-16 15:43:51 -08:00
License Updater
9fa2c4605f licenses: update android licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-01-16 15:43:14 -08:00
Joe Tsai
c25968e1c5 all: make use of ctxkey everywhere (#10846)
Also perform minor cleanups on the ctxkey package itself.
Provide guidance on when to use ctxkey.Key[T] over ctxkey.New.
Also, allow for interface kinds because the value wrapping trick
also happens to fix edge cases with interfaces in Go.

Updates #cleanup

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-01-16 13:56:23 -08:00
Joe Tsai
7732377cd7 tstime/rate: implement Value.{Marshal,Unmarshal}JSON (#8481)
Implement support for marshaling and unmarshaling a Value.

Updates tailscale/corp#8427

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-01-16 13:48:34 -08:00
Irbe Krumina
1c3c3d6752 cmd/k8s-operator: warn if unsupported Ingress Exact path type is used. (#10865)
To reduce the likelihood of breaking users,
if we implement stricter Exact path type matching in the future.

Updates tailscale/tailscale#10730

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-16 17:02:34 +00:00
Irbe Krumina
50b52dbd7d cmd/k8s-operator: sync StatefulSet labels to their Pods (#10861)
So that users have predictable label values to use when configuring network policies.

Updates tailscale/tailscale#10854

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-16 12:51:10 +00:00
Irbe Krumina
d0492fdee5 cmd/k8s-operator: adds a tailscale IngressClass resource, prints warning if class not found. (#10823)
* cmd/k8s-operator/deploy: deploy a Tailscale IngressClass resource.

Some Ingress validating webhooks reject Ingresses with
.spec.ingressClassName for which there is no matching IngressClass.

Additionally, validate that the expected IngressClass is present,
when parsing a tailscale `Ingress`. 
We currently do not utilize the IngressClass,
however we might in the future at which point
we might start requiring that the right class
for this controller instance actually exists.

Updates tailscale/tailscale#10820

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
Co-authored-by: Anton Tolchanov <anton@tailscale.com>
2024-01-16 12:48:15 +00:00
License Updater
381430eeca licenses: update win/apple licenses
Signed-off-by: License Updater <noreply+license-updater@tailscale.com>
2024-01-13 14:02:42 -08:00
Joe Tsai
241a541864 util/ctxkey: add package for type-safe context keys (#10841)
The lack of type-safety in context.WithValue leads to the common pattern
of defining of package-scoped type to ensure global uniqueness:

	type fooKey struct{}

	func withFoo(ctx context, v Foo) context.Context {
		return context.WithValue(ctx, fooKey{}, v)
	}

	func fooValue(ctx context) Foo {
		v, _ := ctx.Value(fooKey{}).(Foo)
		return v
	}

where usage becomes:

	ctx = withFoo(ctx, foo)
	foo := fooValue(ctx)

With many different context keys, this can be quite tedious.

Using generics, we can simplify this as:

	var fooKey = ctxkey.New("mypkg.fooKey", Foo{})

where usage becomes:

	ctx = fooKey.WithValue(ctx, foo)
	foo := fooKey.Value(ctx)

See https://go.dev/issue/49189

Updates #cleanup

Signed-off-by: Joe Tsai <joetsai@digital-static.net>
2024-01-12 17:35:48 -08:00
kari-ts
c9fd166cc6 net/netmon: when a new network is added, trigger netmon update (#10840)
Fixes #10107
2024-01-12 16:03:04 -08:00
Will Norris
236531c5fc ipn/ipnserver: always allow Windows SYSTEM user to connect
When establishing connections to the ipnserver, we validate that the
local user is allowed to connect.  If Tailscale is currently being
managed by a different user (primarily for multi-user Windows installs),
we don't allow the connection.

With the new device web UI, the inbound connection is coming from
tailscaled itself, which is often running as "NT AUTHORITY\SYSTEM".
In this case, we still want to allow the connection, even though it
doesn't match the user running the Tailscale GUI. The SYSTEM user has
full access to everything on the system anyway, so this doesn't escalate
privileges.

Eventually, we want the device web UI to run outside of the tailscaled
process, at which point this exception would probably not be needed.

Updates tailscale/corp#16393

Signed-off-by: Will Norris <will@tailscale.com>
2024-01-12 14:37:53 -08:00
James Tucker
7100b6e721 derp: optimize another per client field alignment
Updates #self

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-12 13:05:39 -08:00
James Tucker
ee20327496 derp: remove unused per-client struct field
Updates #self

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-12 13:05:31 -08:00
OSS Updater
d841ddcb13 go.mod: update web-client-prebuilt module
Signed-off-by: OSS Updater <noreply+oss-updater@tailscale.com>
2024-01-12 16:04:58 -05:00
James Tucker
a7f65b40c5 derp: optimize field order to reduce GC cost
See the field alignment lints for more information.
Reductions are 64->24 and 64->32 respectively.

Updates #self

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-12 13:04:50 -08:00
Charlotte Brandhorst-Satzkorn
e6910974ca cmd/tailscale/cli: add description to exit-node CLI command
This change adds a description to the exit-node CLI command. This
description will be displayed when using `tailscale -h` and `tailscale
exit-node -h`.

Fixes #10787

Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
2024-01-12 10:06:09 -08:00
Irbe Krumina
169778e23b cmd/k8s-operator: minor fix in name gen (#10830)
Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-12 10:08:22 +00:00
Will Norris
b89c113365 client/web: skip connectivity check on https
The manage client always listens on http (non-secure) port 5252.  If the
login client is loaded over https, then the connectivity check to `/ok`
will fail with a mixed-content error. Mixed-content enforcement is a
browser setting that we have no control over, so there's no way around
this.

In this case of the login client being loaded over https, we skip the
connectivity check entirely.  We will always render the sign-in button,
though we don't know for sure if the user has connectivity, so we
provide some additional help text in case they have trouble signing in.

Updates hassio-addons/addon-tailscale#314

Signed-off-by: Will Norris <will@tailscale.com>
2024-01-11 14:51:29 -08:00
James Tucker
ff9c1ebb4a derp: reduce excess goroutines blocking on broadcasts
Observed on one busy derp node, there were 600 goroutines blocked
writing to this channel, which represents not only more blocked routines
than we need, but also excess wake-ups downstream as the latent
goroutines writes represent no new work.

Updates #self

Signed-off-by: James Tucker <james@tailscale.com>
2024-01-11 14:47:17 -08:00
Irbe Krumina
5cc1bfe82d cmd/k8s-operator: remove configuration knob for Connector (#10791)
The configuration knob (that defaulted to Connector being disabled)
was added largely because the Connector CRD had to be installed in a separate step.
Now when the CRD has been added to both chart and static manifest, we can have it on by default.

Updates tailscale/tailscale#10878

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-11 20:03:53 +00:00
Irbe Krumina
469af614b0 cmd/k8s-operator: fix base truncating for extra long Service names (#10825)
cmd/k8s-operator: fix base truncating for extra long Service names

StatefulSet names for ingress/egress proxies are calculated
using Kubernetes name generator and the parent resource name
as a base.
The name generator also cuts the base, but has a higher max cap.
This commit fixes a bug where, if we get a shortened base back
from the generator, we cut off too little as the base that we
have cut will be passed into the generator again, which will
then itself cut less because the base is shorter- so we end up
with a too long name again.

Updates tailscale/tailscale#10807

Co-authored-by: Maisem Ali <maisem@tailscale.com>
Signed-off-by: Irbe Krumina <irbekrm@gmail.com>
2024-01-11 20:02:03 +00:00
Sonia Appasamy
331a6d105f client/web: add initial types for using peer capabilities
Sets up peer capability types for future use within the web client
views and APIs.

Updates tailscale/corp#16695

Signed-off-by: Sonia Appasamy <sonia@tailscale.com>
2024-01-11 11:20:24 -05:00
Andrew Dunham
6540d1f018 wgengine/router: look up absolute path to netsh.exe on Windows
This is in response to logs from a customer that show that we're unable
to run netsh due to the following error:

    router: firewall: adding Tailscale-Process rule to allow UDP for "C:\\Program Files\\Tailscale\\tailscaled.exe" ...
    router: firewall: error adding Tailscale-Process rule: exec: "netsh": cannot run executable found relative to current directory:

There's approximately no reason to ever dynamically look up the path of
a system utility like netsh.exe, so instead let's first look for it
in the System32 directory and only if that fails fall back to the
previous behaviour.

Updates #10804

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I68cfeb4cab091c79ccff3187d35f50359a690573
2024-01-10 20:20:19 -05:00
Irbe Krumina
ca48db0d60 Makefile,build_docker.sh: allow to configure target platform. (#10806)
Build dev tailscale and k8s-operator images for linux/amd64 only by default,
make it possible to configure target build platform via PLATFORM var.

Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-10 19:19:20 +00:00
Flakes Updater
91c7dfe85c go.mod.sri: update SRI hash for go.mod changes
Signed-off-by: Flakes Updater <noreply+flakes-updater@tailscale.com>
2024-01-10 09:25:32 -08:00
Andrew Lytvynov
86e476c8d1 version/mkversion: allow version override with $TS_VERSION_OVERRIDE (#10799)
This is useful to build local binaries with custom versions to test
version-specific logic (like updates).

Updates https://github.com/tailscale/corp/issues/16703

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-01-10 09:03:11 -08:00
Andrew Lytvynov
4ec6a78551 go.mod: update golang-x-crypto fork (#10786)
Pick up a bunch of recent upstream commits.

Updates #8593

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
2024-01-10 09:02:33 -08:00
Will Norris
84ab040f02 safesocket: detect macsys from within tailscaled
Use the helper method from the version package to detect that we are
running the macsys network extension. This method does the same check
for the HOME environment variable (which works fine in most cases) as
well as the name of the executable (which is needed for the web client).

Updates tailscale/corp#16393

Signed-off-by: Will Norris <will@tailscale.com>
2024-01-10 08:15:40 -08:00
OSS Updater
e7d52eb2f8 go.mod: update web-client-prebuilt module
Signed-off-by: OSS Updater <noreply+oss-updater@tailscale.com>
2024-01-10 10:59:53 -05:00
Irbe Krumina
35f49ac99e cmd/k8s-operator: add Connector CRD to Helm chart and static manifests (#10775)
cmd/k8s-operator: add CRD to chart and static manifest

Add functionality to insert CRD to chart at package time.
Insert CRD to static manifests as this is where they are currently consumed from.

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-01-10 14:20:22 +00:00
389 changed files with 24140 additions and 5570 deletions

View File

@@ -47,6 +47,12 @@ jobs:
- name: Checkout repository
uses: actions/checkout@v4
# Install a more recent Go that understands modern go.mod content.
- name: Install Go
uses: actions/setup-go@v4
with:
go-version-file: go.mod
# Initializes the CodeQL tools for scanning.
- name: Initialize CodeQL
uses: github/codeql-action/init@v2

View File

@@ -34,7 +34,7 @@ jobs:
# Note: this is the 'v3' tag as of 2023-08-14
uses: golangci/golangci-lint-action@639cd343e1d3b897ff35927a75193d57cfcba299
with:
version: v1.54.2
version: v1.56
# Show only new issues if it's a pull request.
only-new-issues: true

View File

@@ -2,8 +2,8 @@ name: "Kubernetes manifests"
on:
pull_request:
paths:
- './cmd/k8s-operator/'
- './k8s-operator/'
- 'cmd/k8s-operator/**'
- 'k8s-operator/**'
- '.github/workflows/kubemanifests.yaml'
# Cancel workflow run if there is a newer push to the same PR for which it is

View File

@@ -183,6 +183,19 @@ jobs:
# the equals signs cause great confusion.
run: go test ./... -bench . -benchtime 1x -run "^$"
privileged:
runs-on: ubuntu-22.04
container:
image: golang:latest
options: --privileged
steps:
- name: checkout
uses: actions/checkout@v4
- name: chown
run: chown -R $(id -u):$(id -g) $PWD
- name: privileged tests
run: ./tool/go test ./util/linuxfw
vm:
runs-on: ["self-hosted", "linux", "vm"]
# VM tests run with some privileges, don't let them run on 3p PRs.
@@ -193,9 +206,9 @@ jobs:
- name: Run VM tests
run: ./tool/go test ./tstest/integration/vms -v -no-s3 -run-vm-tests -run=TestRunUbuntu2004
env:
HOME: "/tmp"
HOME: "/var/lib/ghrunner/home"
TMPDIR: "/tmp"
XDB_CACHE_HOME: "/var/lib/ghrunner/cache"
XDG_CACHE_HOME: "/var/lib/ghrunner/cache"
race-build:
runs-on: ubuntu-22.04
@@ -241,9 +254,9 @@ jobs:
goarch: amd64
- goos: openbsd
goarch: amd64
# Plan9
- goos: plan9
goarch: amd64
# Plan9 (disabled until 3p dependencies are fixed)
# - goos: plan9
# goarch: amd64
runs-on: ubuntu-22.04
steps:

View File

@@ -31,7 +31,7 @@
# $ docker exec tailscaled tailscale status
FROM golang:1.21-alpine AS build-env
FROM golang:1.22-alpine AS build-env
WORKDIR /go/src/tailscale

View File

@@ -3,6 +3,8 @@ SYNO_ARCH ?= "amd64"
SYNO_DSM ?= "7"
TAGS ?= "latest"
PLATFORM ?= "flyio" ## flyio==linux/amd64. Set to "" to build all platforms.
vet: ## Run go vet
./tool/go vet ./...
@@ -88,7 +90,7 @@ publishdevimage: ## Build and publish tailscale image to location specified by $
@test "${REPO}" != "ghcr.io/tailscale/tailscale" || (echo "REPO=... must not be ghcr.io/tailscale/tailscale" && exit 1)
@test "${REPO}" != "tailscale/k8s-operator" || (echo "REPO=... must not be tailscale/k8s-operator" && exit 1)
@test "${REPO}" != "ghcr.io/tailscale/k8s-operator" || (echo "REPO=... must not be ghcr.io/tailscale/k8s-operator" && exit 1)
TAGS="${TAGS}" REPOS=${REPO} PUSH=true TARGET=client ./build_docker.sh
TAGS="${TAGS}" REPOS=${REPO} PLATFORM=${PLATFORM} PUSH=true TARGET=client ./build_docker.sh
publishdevoperator: ## Build and publish k8s-operator image to location specified by ${REPO}
@test -n "${REPO}" || (echo "REPO=... required; e.g. REPO=ghcr.io/${USER}/tailscale" && exit 1)
@@ -96,7 +98,7 @@ publishdevoperator: ## Build and publish k8s-operator image to location specifie
@test "${REPO}" != "ghcr.io/tailscale/tailscale" || (echo "REPO=... must not be ghcr.io/tailscale/tailscale" && exit 1)
@test "${REPO}" != "tailscale/k8s-operator" || (echo "REPO=... must not be tailscale/k8s-operator" && exit 1)
@test "${REPO}" != "ghcr.io/tailscale/k8s-operator" || (echo "REPO=... must not be ghcr.io/tailscale/k8s-operator" && exit 1)
TAGS="${TAGS}" REPOS=${REPO} PUSH=true TARGET=operator ./build_docker.sh
TAGS="${TAGS}" REPOS=${REPO} PLATFORM=${PLATFORM} PUSH=true TARGET=operator ./build_docker.sh
help: ## Show this help
@echo "\nSpecify a command. The choices are:\n"

View File

@@ -37,7 +37,7 @@ not open source.
## Building
We always require the latest Go release, currently Go 1.21. (While we build
We always require the latest Go release, currently Go 1.22. (While we build
releases with our [Go fork](https://github.com/tailscale/go/), its use is not
required.)

View File

@@ -1 +1 @@
1.57.0
1.63.0

424
api.md

File diff suppressed because it is too large Load Diff

View File

@@ -10,6 +10,7 @@
package appc
import (
"context"
"net/netip"
"slices"
"strings"
@@ -20,14 +21,19 @@ import (
"tailscale.com/types/logger"
"tailscale.com/types/views"
"tailscale.com/util/dnsname"
"tailscale.com/util/execqueue"
"tailscale.com/util/mak"
)
// RouteAdvertiser is an interface that allows the AppConnector to advertise
// newly discovered routes that need to be served through the AppConnector.
type RouteAdvertiser interface {
// AdvertiseRoute adds a new route advertisement if the route is not already
// being advertised.
AdvertiseRoute(netip.Prefix) error
// AdvertiseRoute adds one or more route advertisements skipping any that
// are already advertised.
AdvertiseRoute(...netip.Prefix) error
// UnadvertiseRoute removes any matching route advertisements.
UnadvertiseRoute(...netip.Prefix) error
}
// AppConnector is an implementation of an AppConnector that performs
@@ -45,12 +51,19 @@ type AppConnector struct {
// mu guards the fields that follow
mu sync.Mutex
// domains is a map of lower case domain names with no trailing dot, to a
// list of resolved IP addresses.
// domains is a map of lower case domain names with no trailing dot, to an
// ordered list of resolved IP addresses.
domains map[string][]netip.Addr
// controlRoutes is the list of routes that were last supplied by control.
controlRoutes []netip.Prefix
// wildcards is the list of domain strings that match subdomains.
wildcards []string
// queue provides ordering for update operations
queue execqueue.ExecQueue
}
// NewAppConnector creates a new AppConnector.
@@ -61,11 +74,33 @@ func NewAppConnector(logf logger.Logf, routeAdvertiser RouteAdvertiser) *AppConn
}
}
// UpdateDomains replaces the current set of configured domains with the
// supplied set of domains. Domains must not contain a trailing dot, and should
// be lower case. If the domain contains a leading '*' label it matches all
// subdomains of a domain.
// UpdateDomainsAndRoutes starts an asynchronous update of the configuration
// given the new domains and routes.
func (e *AppConnector) UpdateDomainsAndRoutes(domains []string, routes []netip.Prefix) {
e.queue.Add(func() {
// Add the new routes first.
e.updateRoutes(routes)
e.updateDomains(domains)
})
}
// UpdateDomains asynchronously replaces the current set of configured domains
// with the supplied set of domains. Domains must not contain a trailing dot,
// and should be lower case. If the domain contains a leading '*' label it
// matches all subdomains of a domain.
func (e *AppConnector) UpdateDomains(domains []string) {
e.queue.Add(func() {
e.updateDomains(domains)
})
}
// Wait waits for the currently scheduled asynchronous configuration changes to
// complete.
func (e *AppConnector) Wait(ctx context.Context) {
e.queue.Wait(ctx)
}
func (e *AppConnector) updateDomains(domains []string) {
e.mu.Lock()
defer e.mu.Unlock()
@@ -97,6 +132,46 @@ func (e *AppConnector) UpdateDomains(domains []string) {
e.logf("handling domains: %v and wildcards: %v", xmaps.Keys(e.domains), e.wildcards)
}
// updateRoutes merges the supplied routes into the currently configured routes. The routes supplied
// by control for UpdateRoutes are supplemental to the routes discovered by DNS resolution, but are
// also more often whole ranges. UpdateRoutes will remove any single address routes that are now
// covered by new ranges.
func (e *AppConnector) updateRoutes(routes []netip.Prefix) {
e.mu.Lock()
defer e.mu.Unlock()
// If there was no change since the last update, no work to do.
if slices.Equal(e.controlRoutes, routes) {
return
}
if err := e.routeAdvertiser.AdvertiseRoute(routes...); err != nil {
e.logf("failed to advertise routes: %v: %v", routes, err)
return
}
var toRemove []netip.Prefix
nextRoute:
for _, r := range routes {
for _, addr := range e.domains {
for _, a := range addr {
if r.Contains(a) && netip.PrefixFrom(a, a.BitLen()) != r {
pfx := netip.PrefixFrom(a, a.BitLen())
toRemove = append(toRemove, pfx)
continue nextRoute
}
}
}
}
if err := e.routeAdvertiser.UnadvertiseRoute(toRemove...); err != nil {
e.logf("failed to unadvertise routes: %v: %v", toRemove, err)
}
e.controlRoutes = routes
}
// Domains returns the currently configured domain list.
func (e *AppConnector) Domains() views.Slice[string] {
e.mu.Lock()
@@ -132,6 +207,16 @@ func (e *AppConnector) ObserveDNSResponse(res []byte) {
return
}
// cnameChain tracks a chain of CNAMEs for a given query in order to reverse
// a CNAME chain back to the original query for flattening. The keys are
// CNAME record targets, and the value is the name the record answers, so
// for www.example.com CNAME example.com, the map would contain
// ["example.com"] = "www.example.com".
var cnameChain map[string]string
// addressRecords is a list of address records found in the response.
var addressRecords map[string][]netip.Addr
for {
h, err := p.AnswerHeader()
if err == dnsmessage.ErrSectionDone {
@@ -147,73 +232,171 @@ func (e *AppConnector) ObserveDNSResponse(res []byte) {
}
continue
}
if h.Type != dnsmessage.TypeA && h.Type != dnsmessage.TypeAAAA {
switch h.Type {
case dnsmessage.TypeCNAME, dnsmessage.TypeA, dnsmessage.TypeAAAA:
default:
if err := p.SkipAnswer(); err != nil {
return
}
continue
}
domain := h.Name.String()
domain := strings.TrimSuffix(strings.ToLower(h.Name.String()), ".")
if len(domain) == 0 {
return
}
domain = strings.TrimSuffix(domain, ".")
domain = strings.ToLower(domain)
e.logf("[v2] observed DNS response for %s", domain)
e.mu.Lock()
addrs, ok := e.domains[domain]
// match wildcard domains
if !ok {
for _, wc := range e.wildcards {
if dnsname.HasSuffix(domain, wc) {
e.domains[domain] = nil
ok = true
break
}
}
}
e.mu.Unlock()
if !ok {
if err := p.SkipAnswer(); err != nil {
return
}
continue
}
var addr netip.Addr
if h.Type == dnsmessage.TypeCNAME {
res, err := p.CNAMEResource()
if err != nil {
return
}
cname := strings.TrimSuffix(strings.ToLower(res.CNAME.String()), ".")
if len(cname) == 0 {
continue
}
mak.Set(&cnameChain, cname, domain)
continue
}
switch h.Type {
case dnsmessage.TypeA:
r, err := p.AResource()
if err != nil {
return
}
addr = netip.AddrFrom4(r.A)
addr := netip.AddrFrom4(r.A)
mak.Set(&addressRecords, domain, append(addressRecords[domain], addr))
case dnsmessage.TypeAAAA:
r, err := p.AAAAResource()
if err != nil {
return
}
addr = netip.AddrFrom16(r.AAAA)
addr := netip.AddrFrom16(r.AAAA)
mak.Set(&addressRecords, domain, append(addressRecords[domain], addr))
default:
if err := p.SkipAnswer(); err != nil {
return
}
continue
}
if slices.Contains(addrs, addr) {
continue
}
if err := e.routeAdvertiser.AdvertiseRoute(netip.PrefixFrom(addr, addr.BitLen())); err != nil {
e.logf("failed to advertise route for %s: %v: %v", domain, addr, err)
continue
}
e.logf("[v2] advertised route for %v: %v", domain, addr)
}
e.mu.Lock()
e.domains[domain] = append(addrs, addr)
e.mu.Unlock()
e.mu.Lock()
defer e.mu.Unlock()
for domain, addrs := range addressRecords {
domain, isRouted := e.findRoutedDomainLocked(domain, cnameChain)
// domain and none of the CNAMEs in the chain are routed
if !isRouted {
continue
}
// advertise each address we have learned for the routed domain, that
// was not already known.
var toAdvertise []netip.Prefix
for _, addr := range addrs {
if !e.isAddrKnownLocked(domain, addr) {
toAdvertise = append(toAdvertise, netip.PrefixFrom(addr, addr.BitLen()))
}
}
e.logf("[v2] observed new routes for %s: %s", domain, toAdvertise)
e.scheduleAdvertisement(domain, toAdvertise...)
}
}
// starting from the given domain that resolved to an address, find it, or any
// of the domains in the CNAME chain toward resolving it, that are routed
// domains, returning the routed domain name and a bool indicating whether a
// routed domain was found.
// e.mu must be held.
func (e *AppConnector) findRoutedDomainLocked(domain string, cnameChain map[string]string) (string, bool) {
var isRouted bool
for {
_, isRouted = e.domains[domain]
if isRouted {
break
}
// match wildcard domains
for _, wc := range e.wildcards {
if dnsname.HasSuffix(domain, wc) {
e.domains[domain] = nil
isRouted = true
break
}
}
next, ok := cnameChain[domain]
if !ok {
break
}
domain = next
}
return domain, isRouted
}
// isAddrKnownLocked returns true if the address is known to be associated with
// the given domain. Known domain tables are updated for covered routes to speed
// up future matches.
// e.mu must be held.
func (e *AppConnector) isAddrKnownLocked(domain string, addr netip.Addr) bool {
if e.hasDomainAddrLocked(domain, addr) {
return true
}
for _, route := range e.controlRoutes {
if route.Contains(addr) {
// record the new address associated with the domain for faster matching in subsequent
// requests and for diagnostic records.
e.addDomainAddrLocked(domain, addr)
return true
}
}
return false
}
// scheduleAdvertisement schedules an advertisement of the given address
// associated with the given domain.
func (e *AppConnector) scheduleAdvertisement(domain string, routes ...netip.Prefix) {
e.queue.Add(func() {
if err := e.routeAdvertiser.AdvertiseRoute(routes...); err != nil {
e.logf("failed to advertise routes for %s: %v: %v", domain, routes, err)
return
}
e.mu.Lock()
defer e.mu.Unlock()
for _, route := range routes {
if !route.IsSingleIP() {
continue
}
addr := route.Addr()
if !e.hasDomainAddrLocked(domain, addr) {
e.addDomainAddrLocked(domain, addr)
e.logf("[v2] advertised route for %v: %v", domain, addr)
}
}
})
}
// hasDomainAddrLocked returns true if the address has been observed in a
// resolution of domain.
func (e *AppConnector) hasDomainAddrLocked(domain string, addr netip.Addr) bool {
_, ok := slices.BinarySearchFunc(e.domains[domain], addr, compareAddr)
return ok
}
// addDomainAddrLocked adds the address to the list of addresses resolved for
// domain and ensures the list remains sorted. Does not attempt to deduplicate.
func (e *AppConnector) addDomainAddrLocked(domain string, addr netip.Addr) {
e.domains[domain] = append(e.domains[domain], addr)
slices.SortFunc(e.domains[domain], compareAddr)
}
func compareAddr(l, r netip.Addr) int {
return l.Compare(r)
}

View File

@@ -4,6 +4,7 @@
package appc
import (
"context"
"net/netip"
"reflect"
"slices"
@@ -11,12 +12,17 @@ import (
xmaps "golang.org/x/exp/maps"
"golang.org/x/net/dns/dnsmessage"
"tailscale.com/appc/appctest"
"tailscale.com/util/mak"
"tailscale.com/util/must"
)
func TestUpdateDomains(t *testing.T) {
ctx := context.Background()
a := NewAppConnector(t.Logf, nil)
a.UpdateDomains([]string{"example.com"})
a.Wait(ctx)
if got, want := a.Domains().AsSlice(), []string{"example.com"}; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
@@ -24,6 +30,7 @@ func TestUpdateDomains(t *testing.T) {
addr := netip.MustParseAddr("192.0.0.8")
a.domains["example.com"] = append(a.domains["example.com"], addr)
a.UpdateDomains([]string{"example.com"})
a.Wait(ctx)
if got, want := a.domains["example.com"], []netip.Addr{addr}; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
@@ -31,16 +38,68 @@ func TestUpdateDomains(t *testing.T) {
// domains are explicitly downcased on set.
a.UpdateDomains([]string{"UP.EXAMPLE.COM"})
a.Wait(ctx)
if got, want := xmaps.Keys(a.domains), []string{"up.example.com"}; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
}
func TestDomainRoutes(t *testing.T) {
rc := &routeCollector{}
func TestUpdateRoutes(t *testing.T) {
ctx := context.Background()
rc := &appctest.RouteCollector{}
a := NewAppConnector(t.Logf, rc)
a.UpdateDomains([]string{"example.com"})
a.updateDomains([]string{"*.example.com"})
// This route should be collapsed into the range
a.ObserveDNSResponse(dnsResponse("a.example.com.", "192.0.2.1"))
a.Wait(ctx)
if !slices.Equal(rc.Routes(), []netip.Prefix{netip.MustParsePrefix("192.0.2.1/32")}) {
t.Fatalf("got %v, want %v", rc.Routes(), []netip.Prefix{netip.MustParsePrefix("192.0.2.1/32")})
}
// This route should not be collapsed or removed
a.ObserveDNSResponse(dnsResponse("b.example.com.", "192.0.0.1"))
a.Wait(ctx)
routes := []netip.Prefix{netip.MustParsePrefix("192.0.2.0/24"), netip.MustParsePrefix("192.0.0.1/32")}
a.updateRoutes(routes)
slices.SortFunc(rc.Routes(), prefixCompare)
rc.SetRoutes(slices.Compact(rc.Routes()))
slices.SortFunc(routes, prefixCompare)
// Ensure that the non-matching /32 is preserved, even though it's in the domains table.
if !slices.EqualFunc(routes, rc.Routes(), prefixEqual) {
t.Errorf("added routes: got %v, want %v", rc.Routes(), routes)
}
// Ensure that the contained /32 is removed, replaced by the /24.
wantRemoved := []netip.Prefix{netip.MustParsePrefix("192.0.2.1/32")}
if !slices.EqualFunc(rc.RemovedRoutes(), wantRemoved, prefixEqual) {
t.Fatalf("unexpected removed routes: %v", rc.RemovedRoutes())
}
}
func TestUpdateRoutesUnadvertisesContainedRoutes(t *testing.T) {
rc := &appctest.RouteCollector{}
a := NewAppConnector(t.Logf, rc)
mak.Set(&a.domains, "example.com", []netip.Addr{netip.MustParseAddr("192.0.2.1")})
rc.SetRoutes([]netip.Prefix{netip.MustParsePrefix("192.0.2.1/32")})
routes := []netip.Prefix{netip.MustParsePrefix("192.0.2.0/24")}
a.updateRoutes(routes)
if !slices.EqualFunc(routes, rc.Routes(), prefixEqual) {
t.Fatalf("got %v, want %v", rc.Routes(), routes)
}
}
func TestDomainRoutes(t *testing.T) {
rc := &appctest.RouteCollector{}
a := NewAppConnector(t.Logf, rc)
a.updateDomains([]string{"example.com"})
a.ObserveDNSResponse(dnsResponse("example.com.", "192.0.0.8"))
a.Wait(context.Background())
want := map[string][]netip.Addr{
"example.com": {netip.MustParseAddr("192.0.0.8")},
@@ -52,51 +111,89 @@ func TestDomainRoutes(t *testing.T) {
}
func TestObserveDNSResponse(t *testing.T) {
rc := &routeCollector{}
ctx := context.Background()
rc := &appctest.RouteCollector{}
a := NewAppConnector(t.Logf, rc)
// a has no domains configured, so it should not advertise any routes
a.ObserveDNSResponse(dnsResponse("example.com.", "192.0.0.8"))
if got, want := rc.routes, ([]netip.Prefix)(nil); !slices.Equal(got, want) {
if got, want := rc.Routes(), ([]netip.Prefix)(nil); !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
wantRoutes := []netip.Prefix{netip.MustParsePrefix("192.0.0.8/32")}
a.UpdateDomains([]string{"example.com"})
a.updateDomains([]string{"example.com"})
a.ObserveDNSResponse(dnsResponse("example.com.", "192.0.0.8"))
if got, want := rc.routes, wantRoutes; !slices.Equal(got, want) {
a.Wait(ctx)
if got, want := rc.Routes(), wantRoutes; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
// a CNAME record chain should result in a route being added if the chain
// matches a routed domain.
a.updateDomains([]string{"www.example.com", "example.com"})
a.ObserveDNSResponse(dnsCNAMEResponse("192.0.0.9", "www.example.com.", "chain.example.com.", "example.com."))
a.Wait(ctx)
wantRoutes = append(wantRoutes, netip.MustParsePrefix("192.0.0.9/32"))
if got, want := rc.Routes(), wantRoutes; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
// a CNAME record chain should result in a route being added if the chain
// even if only found in the middle of the chain
a.ObserveDNSResponse(dnsCNAMEResponse("192.0.0.10", "outside.example.org.", "www.example.com.", "example.org."))
a.Wait(ctx)
wantRoutes = append(wantRoutes, netip.MustParsePrefix("192.0.0.10/32"))
if got, want := rc.Routes(), wantRoutes; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
wantRoutes = append(wantRoutes, netip.MustParsePrefix("2001:db8::1/128"))
a.ObserveDNSResponse(dnsResponse("example.com.", "2001:db8::1"))
if got, want := rc.routes, wantRoutes; !slices.Equal(got, want) {
a.Wait(ctx)
if got, want := rc.Routes(), wantRoutes; !slices.Equal(got, want) {
t.Errorf("got %v; want %v", got, want)
}
// don't re-advertise routes that have already been advertised
a.ObserveDNSResponse(dnsResponse("example.com.", "2001:db8::1"))
if !slices.Equal(rc.routes, wantRoutes) {
t.Errorf("got %v; want %v", rc.routes, wantRoutes)
a.Wait(ctx)
if !slices.Equal(rc.Routes(), wantRoutes) {
t.Errorf("rc.Routes(): got %v; want %v", rc.Routes(), wantRoutes)
}
// don't advertise addresses that are already in a control provided route
pfx := netip.MustParsePrefix("192.0.2.0/24")
a.updateRoutes([]netip.Prefix{pfx})
wantRoutes = append(wantRoutes, pfx)
a.ObserveDNSResponse(dnsResponse("example.com.", "192.0.2.1"))
a.Wait(ctx)
if !slices.Equal(rc.Routes(), wantRoutes) {
t.Errorf("rc.Routes(): got %v; want %v", rc.Routes(), wantRoutes)
}
if !slices.Contains(a.domains["example.com"], netip.MustParseAddr("192.0.2.1")) {
t.Errorf("missing %v from %v", "192.0.2.1", a.domains["exmaple.com"])
}
}
func TestWildcardDomains(t *testing.T) {
rc := &routeCollector{}
ctx := context.Background()
rc := &appctest.RouteCollector{}
a := NewAppConnector(t.Logf, rc)
a.UpdateDomains([]string{"*.example.com"})
a.updateDomains([]string{"*.example.com"})
a.ObserveDNSResponse(dnsResponse("foo.example.com.", "192.0.0.8"))
if got, want := rc.routes, []netip.Prefix{netip.MustParsePrefix("192.0.0.8/32")}; !slices.Equal(got, want) {
a.Wait(ctx)
if got, want := rc.Routes(), []netip.Prefix{netip.MustParsePrefix("192.0.0.8/32")}; !slices.Equal(got, want) {
t.Errorf("routes: got %v; want %v", got, want)
}
if got, want := a.wildcards, []string{"example.com"}; !slices.Equal(got, want) {
t.Errorf("wildcards: got %v; want %v", got, want)
}
a.UpdateDomains([]string{"*.example.com", "example.com"})
a.updateDomains([]string{"*.example.com", "example.com"})
if _, ok := a.domains["foo.example.com"]; !ok {
t.Errorf("expected foo.example.com to be preserved in domains due to wildcard")
}
@@ -105,7 +202,7 @@ func TestWildcardDomains(t *testing.T) {
}
// There was an early regression where the wildcard domain was added repeatedly, this guards against that.
a.UpdateDomains([]string{"*.example.com", "example.com"})
a.updateDomains([]string{"*.example.com", "example.com"})
if len(a.wildcards) != 1 {
t.Errorf("expected only one wildcard domain, got %v", a.wildcards)
}
@@ -148,15 +245,68 @@ func dnsResponse(domain, address string) []byte {
return must.Get(b.Finish())
}
// routeCollector is a test helper that collects the list of routes advertised
type routeCollector struct {
routes []netip.Prefix
func dnsCNAMEResponse(address string, domains ...string) []byte {
addr := netip.MustParseAddr(address)
b := dnsmessage.NewBuilder(nil, dnsmessage.Header{})
b.EnableCompression()
b.StartAnswers()
if len(domains) >= 2 {
for i, domain := range domains[:len(domains)-1] {
b.CNAMEResource(
dnsmessage.ResourceHeader{
Name: dnsmessage.MustNewName(domain),
Type: dnsmessage.TypeCNAME,
Class: dnsmessage.ClassINET,
TTL: 0,
},
dnsmessage.CNAMEResource{
CNAME: dnsmessage.MustNewName(domains[i+1]),
},
)
}
}
domain := domains[len(domains)-1]
switch addr.BitLen() {
case 32:
b.AResource(
dnsmessage.ResourceHeader{
Name: dnsmessage.MustNewName(domain),
Type: dnsmessage.TypeA,
Class: dnsmessage.ClassINET,
TTL: 0,
},
dnsmessage.AResource{
A: addr.As4(),
},
)
case 128:
b.AAAAResource(
dnsmessage.ResourceHeader{
Name: dnsmessage.MustNewName(domain),
Type: dnsmessage.TypeAAAA,
Class: dnsmessage.ClassINET,
TTL: 0,
},
dnsmessage.AAAAResource{
AAAA: addr.As16(),
},
)
default:
panic("invalid address length")
}
return must.Get(b.Finish())
}
// routeCollector implements RouteAdvertiser
var _ RouteAdvertiser = (*routeCollector)(nil)
func (rc *routeCollector) AdvertiseRoute(pfx netip.Prefix) error {
rc.routes = append(rc.routes, pfx)
return nil
func prefixEqual(a, b netip.Prefix) bool {
return a == b
}
func prefixCompare(a, b netip.Prefix) int {
if a.Addr().Compare(b.Addr()) == 0 {
return a.Bits() - b.Bits()
}
return a.Addr().Compare(b.Addr())
}

49
appc/appctest/appctest.go Normal file
View File

@@ -0,0 +1,49 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
package appctest
import (
"net/netip"
"slices"
)
// RouteCollector is a test helper that collects the list of routes advertised
type RouteCollector struct {
routes []netip.Prefix
removedRoutes []netip.Prefix
}
func (rc *RouteCollector) AdvertiseRoute(pfx ...netip.Prefix) error {
rc.routes = append(rc.routes, pfx...)
return nil
}
func (rc *RouteCollector) UnadvertiseRoute(toRemove ...netip.Prefix) error {
routes := rc.routes
rc.routes = rc.routes[:0]
for _, r := range routes {
if !slices.Contains(toRemove, r) {
rc.routes = append(rc.routes, r)
} else {
rc.removedRoutes = append(rc.removedRoutes, r)
}
}
return nil
}
// RemovedRoutes returns the list of routes that were removed.
func (rc *RouteCollector) RemovedRoutes() []netip.Prefix {
return rc.removedRoutes
}
// Routes returns the ordered list of routes that were added, including
// possible duplicates.
func (rc *RouteCollector) Routes() []netip.Prefix {
return rc.routes
}
func (rc *RouteCollector) SetRoutes(routes []netip.Prefix) error {
rc.routes = routes
return nil
}

View File

@@ -20,9 +20,9 @@
set -eu
# Use the "go" binary from the "tool" directory (which is github.com/tailscale/go)
export PATH=$PWD/tool:$PATH
export PATH="$PWD"/tool:"$PATH"
eval $(./build_dist.sh shellvars)
eval "$(./build_dist.sh shellvars)"
DEFAULT_TARGET="client"
DEFAULT_TAGS="v${VERSION_SHORT},v${VERSION_MINOR}"
@@ -32,6 +32,7 @@ PUSH="${PUSH:-false}"
TARGET="${TARGET:-${DEFAULT_TARGET}}"
TAGS="${TAGS:-${DEFAULT_TAGS}}"
BASE="${BASE:-${DEFAULT_BASE}}"
PLATFORM="${PLATFORM:-}" # default to all platforms
case "$TARGET" in
client)
@@ -50,6 +51,7 @@ case "$TARGET" in
--tags="${TAGS}" \
--repos="${REPOS}" \
--push="${PUSH}" \
--target="${PLATFORM}" \
/usr/local/bin/containerboot
;;
operator)
@@ -65,10 +67,11 @@ case "$TARGET" in
--tags="${TAGS}" \
--repos="${REPOS}" \
--push="${PUSH}" \
--target="${PLATFORM}" \
/usr/local/bin/operator
;;
*)
echo "unknown target: $TARGET"
exit 1
;;
esac
esac

View File

@@ -270,6 +270,14 @@ type UserRuleMatch struct {
Users []string `json:"users"`
Ports []string `json:"ports"`
LineNumber int `json:"lineNumber"`
// Postures is a list of posture policies that are
// associated with this match. The rules can be looked
// up in the ACLPreviewResponse parent struct.
// The source of the list is from srcPosture on
// an ACL or Grant rule:
// https://tailscale.com/kb/1288/device-posture#posture-conditions
Postures []string `json:"postures"`
}
// ACLPreviewResponse is the response type of previewACLPostRequest
@@ -277,6 +285,12 @@ type ACLPreviewResponse struct {
Matches []UserRuleMatch `json:"matches"` // ACL rules that match the specified user or ipport.
Type string `json:"type"` // The request type: currently only "user" or "ipport".
PreviewFor string `json:"previewFor"` // A specific user or ipport.
// Postures is a map of postures and associated rules that apply
// to this preview.
// For more details about the posture mapping, see:
// https://tailscale.com/kb/1288/device-posture#postures
Postures map[string][]string `json:"postures,omitempty"`
}
// ACLPreview is the response type of PreviewACLForUser, PreviewACLForIPPort, PreviewACLHuJSONForUser, and PreviewACLHuJSONForIPPort
@@ -284,6 +298,12 @@ type ACLPreview struct {
Matches []UserRuleMatch `json:"matches"`
User string `json:"user,omitempty"` // Filled if response of PreviewACLForUser or PreviewACLHuJSONForUser
IPPort string `json:"ipport,omitempty"` // Filled if response of PreviewACLForIPPort or PreviewACLHuJSONForIPPort
// Postures is a map of postures and associated rules that apply
// to this preview.
// For more details about the posture mapping, see:
// https://tailscale.com/kb/1288/device-posture#postures
Postures map[string][]string `json:"postures,omitempty"`
}
func (c *Client) previewACLPostRequest(ctx context.Context, body []byte, previewType string, previewFor string) (res *ACLPreviewResponse, err error) {
@@ -341,8 +361,9 @@ func (c *Client) PreviewACLForUser(ctx context.Context, acl ACL, user string) (r
}
return &ACLPreview{
Matches: b.Matches,
User: b.PreviewFor,
Matches: b.Matches,
User: b.PreviewFor,
Postures: b.Postures,
}, nil
}
@@ -369,8 +390,9 @@ func (c *Client) PreviewACLForIPPort(ctx context.Context, acl ACL, ipport netip.
}
return &ACLPreview{
Matches: b.Matches,
IPPort: b.PreviewFor,
Matches: b.Matches,
IPPort: b.PreviewFor,
Postures: b.Postures,
}, nil
}
@@ -394,8 +416,9 @@ func (c *Client) PreviewACLHuJSONForUser(ctx context.Context, acl ACLHuJSON, use
}
return &ACLPreview{
Matches: b.Matches,
User: b.PreviewFor,
Matches: b.Matches,
User: b.PreviewFor,
Postures: b.Postures,
}, nil
}
@@ -419,8 +442,9 @@ func (c *Client) PreviewACLHuJSONForIPPort(ctx context.Context, acl ACLHuJSON, i
}
return &ACLPreview{
Matches: b.Matches,
IPPort: b.PreviewFor,
Matches: b.Matches,
IPPort: b.PreviewFor,
Postures: b.Postures,
}, nil
}

View File

@@ -7,6 +7,7 @@ package tailscale
import (
"bytes"
"cmp"
"context"
"crypto/tls"
"encoding/json"
@@ -34,10 +35,10 @@ import (
"tailscale.com/paths"
"tailscale.com/safesocket"
"tailscale.com/tailcfg"
"tailscale.com/tailfs"
"tailscale.com/tka"
"tailscale.com/types/key"
"tailscale.com/types/tkatype"
"tailscale.com/util/cmpx"
)
// defaultLocalClient is the default LocalClient when using the legacy
@@ -479,7 +480,7 @@ func (lc *LocalClient) DebugPortmap(ctx context.Context, opts *DebugPortmapOpts)
opts = &DebugPortmapOpts{}
}
vals.Set("duration", cmpx.Or(opts.Duration, 5*time.Second).String())
vals.Set("duration", cmp.Or(opts.Duration, 5*time.Second).String())
vals.Set("type", opts.Type)
vals.Set("log_http", strconv.FormatBool(opts.LogHTTP))
@@ -1417,6 +1418,57 @@ func (lc *LocalClient) CheckUpdate(ctx context.Context) (*tailcfg.ClientVersion,
return &cv, nil
}
// TailFSSetFileServerAddr instructs TailFS to use the server at addr to access
// the filesystem. This is used on platforms like Windows and MacOS to let
// TailFS know to use the file server running in the GUI app.
func (lc *LocalClient) TailFSSetFileServerAddr(ctx context.Context, addr string) error {
_, err := lc.send(ctx, "PUT", "/localapi/v0/tailfs/fileserver-address", http.StatusCreated, strings.NewReader(addr))
return err
}
// TailFSShareSet adds or updates the given share in the list of shares that
// TailFS will serve to remote nodes. If a share with the same name already
// exists, the existing share is replaced/updated.
func (lc *LocalClient) TailFSShareSet(ctx context.Context, share *tailfs.Share) error {
_, err := lc.send(ctx, "PUT", "/localapi/v0/tailfs/shares", http.StatusCreated, jsonBody(share))
return err
}
// TailFSShareRemove removes the share with the given name from the list of
// shares that TailFS will serve to remote nodes.
func (lc *LocalClient) TailFSShareRemove(ctx context.Context, name string) error {
_, err := lc.send(
ctx,
"DELETE",
"/localapi/v0/tailfs/shares",
http.StatusNoContent,
strings.NewReader(name))
return err
}
// TailFSShareRename renames the share from old to new name.
func (lc *LocalClient) TailFSShareRename(ctx context.Context, oldName, newName string) error {
_, err := lc.send(
ctx,
"POST",
"/localapi/v0/tailfs/shares",
http.StatusNoContent,
jsonBody([2]string{oldName, newName}))
return err
}
// TailFSShareList returns the list of shares that TailFS is currently serving
// to remote nodes.
func (lc *LocalClient) TailFSShareList(ctx context.Context) ([]*tailfs.Share, error) {
result, err := lc.get200(ctx, "/localapi/v0/tailfs/shares")
if err != nil {
return nil, err
}
var shares []*tailfs.Share
err = json.Unmarshal(result, &shares)
return shares, err
}
// IPNBusWatcher is an active subscription (watch) of the local tailscaled IPN bus.
// It's returned by LocalClient.WatchIPNBus.
//

View File

@@ -5,7 +5,11 @@
package tailscale
import "testing"
import (
"testing"
"tailscale.com/tstest/deptest"
)
func TestGetServeConfigFromJSON(t *testing.T) {
sc, err := getServeConfigFromJSON([]byte("null"))
@@ -25,3 +29,14 @@ func TestGetServeConfigFromJSON(t *testing.T) {
t.Errorf("want non-nil TCP for object")
}
}
func TestDeps(t *testing.T) {
deptest.DepChecker{
BadDeps: map[string]string{
// Make sure we don't again accidentally bring in a dependency on
// TailFS or its transitive dependencies
"tailscale.com/tailfs/tailfsimpl": "https://github.com/tailscale/tailscale/pull/10631",
"github.com/studio-b12/gowebdav": "https://github.com/tailscale/tailscale/pull/10631",
},
}.Check(t)
}

View File

@@ -8,8 +8,10 @@ import (
"crypto/rand"
"encoding/base64"
"errors"
"fmt"
"net/http"
"net/url"
"slices"
"strings"
"time"
@@ -232,3 +234,106 @@ func (s *Server) newSessionID() (string, error) {
}
return "", errors.New("too many collisions generating new session; please refresh page")
}
// peerCapabilities holds information about what a source
// peer is allowed to edit via the web UI.
//
// map value is true if the peer can edit the given feature.
// Only capFeatures included in validCaps will be included.
type peerCapabilities map[capFeature]bool
// canEdit is true if the peerCapabilities grant edit access
// to the given feature.
func (p peerCapabilities) canEdit(feature capFeature) bool {
if p == nil {
return false
}
if p[capFeatureAll] {
return true
}
return p[feature]
}
// isEmpty is true if p is either nil or has no capabilities
// with value true.
func (p peerCapabilities) isEmpty() bool {
if p == nil {
return true
}
for _, v := range p {
if v == true {
return false
}
}
return true
}
type capFeature string
const (
// The following values should not be edited.
// New caps can be added, but existing ones should not be changed,
// as these exact values are used by users in tailnet policy files.
//
// IMPORTANT: When adding a new cap, also update validCaps slice below.
capFeatureAll capFeature = "*" // grants peer management of all features
capFeatureSSH capFeature = "ssh" // grants peer SSH server management
capFeatureSubnets capFeature = "subnets" // grants peer subnet routes management
capFeatureExitNodes capFeature = "exitnodes" // grants peer ability to advertise-as and use exit nodes
capFeatureAccount capFeature = "account" // grants peer ability to turn on auto updates and log out of node
)
// validCaps contains the list of valid capabilities used in the web client.
// Any capabilities included in a peer's grants that do not fall into this
// list will be ignored.
var validCaps []capFeature = []capFeature{
capFeatureAll,
capFeatureSSH,
capFeatureSubnets,
capFeatureExitNodes,
capFeatureAccount,
}
type capRule struct {
CanEdit []string `json:"canEdit,omitempty"` // list of features peer is allowed to edit
}
// toPeerCapabilities parses out the web ui capabilities from the
// given whois response.
func toPeerCapabilities(status *ipnstate.Status, whois *apitype.WhoIsResponse) (peerCapabilities, error) {
if whois == nil || status == nil {
return peerCapabilities{}, nil
}
if whois.Node.IsTagged() {
// We don't allow management *from* tagged nodes, so ignore caps.
// The web client auth flow relies on having a true user identity
// that can be verified through login.
return peerCapabilities{}, nil
}
if !status.Self.IsTagged() {
// User owned nodes are only ever manageable by the owner.
if status.Self.UserID != whois.UserProfile.ID {
return peerCapabilities{}, nil
} else {
return peerCapabilities{capFeatureAll: true}, nil // owner can edit all features
}
}
// For tagged nodes, we actually look at the granted capabilities.
caps := peerCapabilities{}
rules, err := tailcfg.UnmarshalCapJSON[capRule](whois.CapMap, tailcfg.PeerCapabilityWebUI)
if err != nil {
return nil, fmt.Errorf("failed to unmarshal capability: %v", err)
}
for _, c := range rules {
for _, f := range c.CanEdit {
cap := capFeature(strings.ToLower(f))
if slices.Contains(validCaps, cap) {
caps[cap] = true
}
}
}
return caps, nil
}

View File

@@ -6,6 +6,7 @@
"node": "18.16.1",
"yarn": "1.22.19"
},
"type": "module",
"private": true,
"dependencies": {
"@radix-ui/react-collapsible": "^1.0.3",
@@ -19,23 +20,28 @@
"zustand": "^4.4.7"
},
"devDependencies": {
"@types/node": "^18.16.1",
"@types/react": "^18.0.20",
"@types/react-dom": "^18.0.6",
"@vitejs/plugin-react-swc": "^3.3.2",
"@vitejs/plugin-react-swc": "^3.6.0",
"autoprefixer": "^10.4.15",
"eslint": "^8.23.1",
"eslint-config-react-app": "^7.0.1",
"eslint-plugin-curly-quotes": "^1.0.4",
"jsdom": "^23.0.1",
"postcss": "^8.4.31",
"prettier": "^2.5.1",
"prettier-plugin-organize-imports": "^3.2.2",
"tailwindcss": "^3.3.3",
"typescript": "^4.7.4",
"vite": "^4.3.9",
"vite-plugin-rewrite-all": "^1.0.1",
"vite-plugin-svgr": "^3.2.0",
"typescript": "^5.3.3",
"vite": "^5.1.4",
"vite-plugin-svgr": "^4.2.0",
"vite-tsconfig-paths": "^3.5.0",
"vitest": "^0.32.0"
"vitest": "^1.3.1"
},
"resolutions": {
"@typescript-eslint/eslint-plugin": "^6.2.1",
"@typescript-eslint/parser": "^6.2.1"
},
"scripts": {
"build": "vite build",
@@ -50,9 +56,11 @@
"react-app"
],
"plugins": [
"curly-quotes",
"react-hooks"
],
"rules": {
"curly-quotes/no-straight-quotes": "warn",
"react-hooks/rules-of-hooks": "error",
"react-hooks/exhaustive-deps": "error"
},

View File

@@ -4,8 +4,8 @@
import * as Primitive from "@radix-ui/react-popover"
import cx from "classnames"
import React, { useCallback } from "react"
import { ReactComponent as ChevronDown } from "src/assets/icons/chevron-down.svg"
import { ReactComponent as Copy } from "src/assets/icons/copy.svg"
import ChevronDown from "src/assets/icons/chevron-down.svg?react"
import Copy from "src/assets/icons/copy.svg?react"
import NiceIP from "src/components/nice-ip"
import useToaster from "src/hooks/toaster"
import Button from "src/ui/button"

View File

@@ -2,7 +2,7 @@
// SPDX-License-Identifier: BSD-3-Clause
import React from "react"
import { ReactComponent as TailscaleIcon } from "src/assets/icons/tailscale-icon.svg"
import TailscaleIcon from "src/assets/icons/tailscale-icon.svg?react"
import LoginToggle from "src/components/login-toggle"
import DeviceDetailsView from "src/components/views/device-details-view"
import DisconnectedView from "src/components/views/disconnected-view"
@@ -11,8 +11,8 @@ import LoginView from "src/components/views/login-view"
import SSHView from "src/components/views/ssh-view"
import SubnetRouterView from "src/components/views/subnet-router-view"
import { UpdatingView } from "src/components/views/updating-view"
import useAuth, { AuthResponse } from "src/hooks/auth"
import { Feature, featureDescription, NodeData } from "src/types"
import useAuth, { AuthResponse, canEdit } from "src/hooks/auth"
import { Feature, NodeData, featureDescription } from "src/types"
import Card from "src/ui/card"
import EmptyState from "src/ui/empty-state"
import LoadingDots from "src/ui/loading-dots"
@@ -56,16 +56,19 @@ function WebClient({
<Header node={node} auth={auth} newSession={newSession} />
<Switch>
<Route path="/">
<HomeView readonly={!auth.canManageNode} node={node} />
<HomeView node={node} auth={auth} />
</Route>
<Route path="/details">
<DeviceDetailsView readonly={!auth.canManageNode} node={node} />
<DeviceDetailsView node={node} auth={auth} />
</Route>
<FeatureRoute path="/subnets" feature="advertise-routes" node={node}>
<SubnetRouterView readonly={!auth.canManageNode} node={node} />
<SubnetRouterView
readonly={!canEdit("subnets", auth)}
node={node}
/>
</FeatureRoute>
<FeatureRoute path="/ssh" feature="ssh" node={node}>
<SSHView readonly={!auth.canManageNode} node={node} />
<SSHView readonly={!canEdit("ssh", auth)} node={node} />
</FeatureRoute>
{/* <Route path="/serve">Share local content</Route> */}
<FeatureRoute path="/update" feature="auto-update" node={node}>

View File

@@ -4,8 +4,8 @@
import cx from "classnames"
import React, { useCallback, useEffect, useMemo, useRef, useState } from "react"
import { useAPI } from "src/api"
import { ReactComponent as Check } from "src/assets/icons/check.svg"
import { ReactComponent as ChevronDown } from "src/assets/icons/chevron-down.svg"
import Check from "src/assets/icons/check.svg?react"
import ChevronDown from "src/assets/icons/chevron-down.svg?react"
import useExitNodes, {
noExitNode,
runAsExitNode,
@@ -180,7 +180,7 @@ export default function ExitNodeSelector({
)}
{pending && (
<p className="text-white p-3">
Pending approval to run as exit node. This device won't be usable as
Pending approval to run as exit node. This device wont be usable as
an exit node until then.
</p>
)}

View File

@@ -2,15 +2,17 @@
// SPDX-License-Identifier: BSD-3-Clause
import cx from "classnames"
import React, { useCallback, useEffect, useState } from "react"
import { ReactComponent as ChevronDown } from "src/assets/icons/chevron-down.svg"
import { ReactComponent as Eye } from "src/assets/icons/eye.svg"
import { ReactComponent as User } from "src/assets/icons/user.svg"
import { AuthResponse, AuthType } from "src/hooks/auth"
import React, { useCallback, useMemo, useState } from "react"
import ChevronDown from "src/assets/icons/chevron-down.svg?react"
import Eye from "src/assets/icons/eye.svg?react"
import User from "src/assets/icons/user.svg?react"
import { AuthResponse, hasAnyEditCapabilities } from "src/hooks/auth"
import { useTSWebConnected } from "src/hooks/ts-web-connected"
import { NodeData } from "src/types"
import Button from "src/ui/button"
import Popover from "src/ui/popover"
import ProfilePic from "src/ui/profile-pic"
import { assertNever, isHTTPS } from "src/utils/util"
export default function LoginToggle({
node,
@@ -22,12 +24,29 @@ export default function LoginToggle({
newSession: () => Promise<void>
}) {
const [open, setOpen] = useState<boolean>(false)
const { tsWebConnected, checkTSWebConnection } = useTSWebConnected(
auth.serverMode,
node.IPv4
)
return (
<Popover
className="p-3 bg-white rounded-lg shadow flex flex-col gap-2 max-w-[317px]"
className="p-3 bg-white rounded-lg shadow flex flex-col max-w-[317px]"
content={
<LoginPopoverContent node={node} auth={auth} newSession={newSession} />
auth.serverMode === "readonly" ? (
<ReadonlyModeContent auth={auth} />
) : auth.serverMode === "login" ? (
<LoginModeContent
auth={auth}
node={node}
tsWebConnected={tsWebConnected}
checkTSWebConnection={checkTSWebConnection}
/>
) : auth.serverMode === "manage" ? (
<ManageModeContent auth={auth} node={node} newSession={newSession} />
) : (
assertNever(auth.serverMode)
)
}
side="bottom"
align="end"
@@ -35,201 +54,303 @@ export default function LoginToggle({
onOpenChange={setOpen}
asChild
>
{!auth.canManageNode ? (
<button
className={cx(
"pl-3 py-1 bg-gray-700 rounded-full flex justify-start items-center h-[34px]",
{ "pr-1": auth.viewerIdentity, "pr-3": !auth.viewerIdentity }
)}
onClick={() => setOpen(!open)}
>
<Eye />
<div className="text-white leading-snug ml-2 mr-1">Viewing</div>
<ChevronDown className="stroke-white w-[15px] h-[15px]" />
{auth.viewerIdentity && (
<ProfilePic
className="ml-2"
size="medium"
url={auth.viewerIdentity.profilePicUrl}
/>
)}
</button>
) : (
<div
className={cx(
"w-[34px] h-[34px] p-1 rounded-full justify-center items-center inline-flex hover:bg-gray-300",
{
"bg-transparent": !open,
"bg-gray-300": open,
}
)}
>
<button onClick={() => setOpen(!open)}>
<ProfilePic
size="medium"
url={auth.viewerIdentity?.profilePicUrl}
/>
</button>
</div>
)}
<div>
{auth.authorized ? (
<TriggerWhenManaging auth={auth} open={open} setOpen={setOpen} />
) : (
<TriggerWhenReading auth={auth} open={open} setOpen={setOpen} />
)}
</div>
</Popover>
)
}
function LoginPopoverContent({
/**
* TriggerWhenManaging is displayed as the trigger for the login popover
* when the user has an active authorized managment session.
*/
function TriggerWhenManaging({
auth,
open,
setOpen,
}: {
auth: AuthResponse
open: boolean
setOpen: (next: boolean) => void
}) {
return (
<div
className={cx(
"w-[34px] h-[34px] p-1 rounded-full justify-center items-center inline-flex hover:bg-gray-300",
{
"bg-transparent": !open,
"bg-gray-300": open,
}
)}
>
<button onClick={() => setOpen(!open)}>
<ProfilePic size="medium" url={auth.viewerIdentity?.profilePicUrl} />
</button>
</div>
)
}
/**
* TriggerWhenReading is displayed as the trigger for the login popover
* when the user is currently in read mode (doesn't have an authorized
* management session).
*/
function TriggerWhenReading({
auth,
open,
setOpen,
}: {
auth: AuthResponse
open: boolean
setOpen: (next: boolean) => void
}) {
return (
<button
className={cx(
"pl-3 py-1 bg-gray-700 rounded-full flex justify-start items-center h-[34px]",
{ "pr-1": auth.viewerIdentity, "pr-3": !auth.viewerIdentity }
)}
onClick={() => setOpen(!open)}
>
<Eye />
<div className="text-white leading-snug ml-2 mr-1">Viewing</div>
<ChevronDown className="stroke-white w-[15px] h-[15px]" />
{auth.viewerIdentity && (
<ProfilePic
className="ml-2"
size="medium"
url={auth.viewerIdentity.profilePicUrl}
/>
)}
</button>
)
}
/**
* PopoverContentHeader is the header for the login popover.
*/
function PopoverContentHeader({ auth }: { auth: AuthResponse }) {
return (
<div className="text-black text-sm font-medium leading-tight mb-1">
{auth.authorized ? "Managing" : "Viewing"}
{auth.viewerIdentity && ` as ${auth.viewerIdentity.loginName}`}
</div>
)
}
/**
* PopoverContentFooter is the footer for the login popover.
*/
function PopoverContentFooter({ auth }: { auth: AuthResponse }) {
return auth.viewerIdentity ? (
<>
<hr className="my-2" />
<div className="flex items-center">
<User className="flex-shrink-0" />
<p className="text-gray-500 text-xs ml-2">
We recognize you because you are accessing this page from{" "}
<span className="font-medium">
{auth.viewerIdentity.nodeName || auth.viewerIdentity.nodeIP}
</span>
</p>
</div>
</>
) : null
}
/**
* ReadonlyModeContent is the body of the login popover when the web
* client is being run in "readonly" server mode.
*/
function ReadonlyModeContent({ auth }: { auth: AuthResponse }) {
return (
<>
<PopoverContentHeader auth={auth} />
<p className="text-gray-500 text-xs">
This web interface is running in read-only mode.{" "}
<a
href="https://tailscale.com/s/web-client-read-only"
className="text-blue-700"
target="_blank"
rel="noreferrer"
>
Learn more &rarr;
</a>
</p>
<PopoverContentFooter auth={auth} />
</>
)
}
/**
* LoginModeContent is the body of the login popover when the web
* client is being run in "login" server mode.
*/
function LoginModeContent({
node,
auth,
tsWebConnected,
checkTSWebConnection,
}: {
node: NodeData
auth: AuthResponse
tsWebConnected: boolean
checkTSWebConnection: () => void
}) {
const https = isHTTPS()
// We can't run the ts web connection test when the webpage is loaded
// over HTTPS. So in this case, we default to presenting a login button
// with some helper text reminding the user to check their connection
// themselves.
const hasACLAccess = https || tsWebConnected
const hasEditCaps = useMemo(() => {
if (!auth.viewerIdentity) {
// If not connected to login client over tailscale, we won't know the viewer's
// identity. So we must assume they may be able to edit something and have the
// management client handle permissions once the user gets there.
return true
}
return hasAnyEditCapabilities(auth)
}, [auth])
const handleLogin = useCallback(() => {
// Must be connected over Tailscale to log in.
// Send user to Tailscale IP and start check mode
const manageURL = `http://${node.IPv4}:5252/?check=now`
if (window.self !== window.top) {
// If we're inside an iframe, open management client in new window.
window.open(manageURL, "_blank")
} else {
window.location.href = manageURL
}
}, [node.IPv4])
return (
<div
onMouseEnter={
hasEditCaps && !hasACLAccess ? checkTSWebConnection : undefined
}
>
<PopoverContentHeader auth={auth} />
{!hasACLAccess || !hasEditCaps ? (
<>
<p className="text-gray-500 text-xs">
{!hasEditCaps ? (
// ACLs allow access, but user isn't allowed to edit any features,
// restricted to readonly. No point in sending them over to the
// tailscaleIP:5252 address.
<>
You dont have permission to make changes to this device, but
you can view most of its details.
</>
) : !node.ACLAllowsAnyIncomingTraffic ? (
// Tailnet ACLs don't allow access to anyone.
<>
The current tailnet policy file does not allow connecting to
this device.
</>
) : (
// ACLs don't allow access to this user specifically.
<>
Cannot access this devices Tailscale IP. Make sure you are
connected to your tailnet, and that your policy file allows
access.
</>
)}{" "}
<a
href="https://tailscale.com/s/web-client-access"
className="text-blue-700"
target="_blank"
rel="noreferrer"
>
Learn more &rarr;
</a>
</p>
</>
) : (
// User can connect to Tailcale IP; sign in when ready.
<>
<p className="text-gray-500 text-xs">
You can see most of this devices details. To make changes, you need
to sign in.
</p>
{https && (
// we don't know if the user can connect over TS, so
// provide extra tips in case they have trouble.
<p className="text-gray-500 text-xs font-semibold pt-2">
Make sure you are connected to your tailnet, and that your policy
file allows access.
</p>
)}
<SignInButton auth={auth} onClick={handleLogin} />
</>
)}
<PopoverContentFooter auth={auth} />
</div>
)
}
/**
* ManageModeContent is the body of the login popover when the web
* client is being run in "manage" server mode.
*/
function ManageModeContent({
auth,
newSession,
}: {
node: NodeData
auth: AuthResponse
newSession: () => Promise<void>
newSession: () => void
}) {
/**
* canConnectOverTS indicates whether the current viewer
* is able to hit the node's web client that's being served
* at http://${node.IP}:5252. If false, this means that the
* viewer must connect to the correct tailnet before being
* able to sign in.
*/
const [canConnectOverTS, setCanConnectOverTS] = useState<boolean>(false)
const [isRunningCheck, setIsRunningCheck] = useState<boolean>(false)
const checkTSConnection = useCallback(() => {
if (auth.viewerIdentity) {
setCanConnectOverTS(true) // already connected over ts
return
}
// Otherwise, test connection to the ts IP.
if (isRunningCheck) {
return // already checking
}
setIsRunningCheck(true)
fetch(`http://${node.IPv4}:5252/ok`, { mode: "no-cors" })
.then(() => {
setCanConnectOverTS(true)
setIsRunningCheck(false)
})
.catch(() => setIsRunningCheck(false))
}, [auth.viewerIdentity, isRunningCheck, node.IPv4])
/**
* Checking connection for first time on page load.
*
* While not connected, we check again whenever the mouse
* enters the popover component, to pick up on the user
* leaving to turn on Tailscale then returning to the view.
* See `onMouseEnter` on the div below.
*/
// eslint-disable-next-line react-hooks/exhaustive-deps
useEffect(() => checkTSConnection(), [])
const handleSignInClick = useCallback(() => {
if (auth.viewerIdentity && auth.serverMode === "manage") {
if (window.self !== window.top) {
// if we're inside an iframe, start session in new window
let url = new URL(window.location.href)
url.searchParams.set("check", "now")
window.open(url, "_blank")
} else {
newSession()
}
const handleLogin = useCallback(() => {
if (window.self !== window.top) {
// If we're inside an iframe, start session in new window.
let url = new URL(window.location.href)
url.searchParams.set("check", "now")
window.open(url, "_blank")
} else {
// Must be connected over Tailscale to log in.
// Send user to Tailscale IP and start check mode
const manageURL = `http://${node.IPv4}:5252/?check=now`
if (window.self !== window.top) {
// if we're inside an iframe, open management client in new window
window.open(manageURL, "_blank")
} else {
window.location.href = manageURL
}
newSession()
}
}, [auth.viewerIdentity, auth.serverMode, newSession, node.IPv4])
}, [newSession])
const hasAnyPermissions = useMemo(() => hasAnyEditCapabilities(auth), [auth])
return (
<div onMouseEnter={!canConnectOverTS ? checkTSConnection : undefined}>
<div className="text-black text-sm font-medium leading-tight mb-1">
{!auth.canManageNode ? "Viewing" : "Managing"}
{auth.viewerIdentity && ` as ${auth.viewerIdentity.loginName}`}
</div>
{!auth.canManageNode && (
<>
{!auth.viewerIdentity ? (
// User is not connected over Tailscale.
// These states are only possible on the login client.
<>
{!canConnectOverTS ? (
<>
<p className="text-gray-500 text-xs">
{!node.ACLAllowsAnyIncomingTraffic ? (
// Tailnet ACLs don't allow access.
<>
The current tailnet policy file does not allow
connecting to this device.
</>
) : (
// ACLs allow access, but user can't connect.
<>
Cannot access this device's Tailscale IP. Make sure you
are connected to your tailnet, and that your policy file
allows access.
</>
)}{" "}
<a
href="https://tailscale.com/s/web-client-connection"
className="text-blue-700"
target="_blank"
rel="noreferrer"
>
Learn more &rarr;
</a>
</p>
</>
) : (
// User can connect to Tailcale IP; sign in when ready.
<>
<p className="text-gray-500 text-xs">
You can see most of this device's details. To make changes,
you need to sign in.
</p>
<SignInButton auth={auth} onClick={handleSignInClick} />
</>
)}
</>
) : auth.authNeeded === AuthType.tailscale ? (
// User is connected over Tailscale, but needs to complete check mode.
<>
<p className="text-gray-500 text-xs">
To make changes, sign in to confirm your identity. This extra
step helps us keep your device secure.
</p>
<SignInButton auth={auth} onClick={handleSignInClick} />
</>
) : (
// User is connected over tailscale, but doesn't have permission to manage.
<>
<PopoverContentHeader auth={auth} />
{!auth.authorized &&
(hasAnyPermissions ? (
// User is connected over Tailscale, but needs to complete check mode.
<>
<p className="text-gray-500 text-xs">
You dont have permission to make changes to this device, but you
can view most of its details.
To make changes, sign in to confirm your identity. This extra step
helps us keep your device secure.
</p>
)}
</>
)}
{auth.viewerIdentity && (
<>
<hr className="my-2" />
<div className="flex items-center">
<User className="flex-shrink-0" />
<p className="text-gray-500 text-xs ml-2">
We recognize you because you are accessing this page from{" "}
<span className="font-medium">
{auth.viewerIdentity.nodeName || auth.viewerIdentity.nodeIP}
</span>
</p>
</div>
</>
)}
</div>
<SignInButton auth={auth} onClick={handleLogin} />
</>
) : (
// User is connected over tailscale, but doesn't have permission to manage.
<p className="text-gray-500 text-xs">
You dont have permission to make changes to this device, but you
can view most of its details.{" "}
<a
href="https://tailscale.com/s/web-client-access"
className="text-blue-700"
target="_blank"
rel="noreferrer"
>
Learn more &rarr;
</a>
</p>
))}
<PopoverContentFooter auth={auth} />
</>
)
}

View File

@@ -61,7 +61,7 @@ export function ChangelogText({ version }: { version?: string }) {
<a href="https://tailscale.com/changelog/" className="link">
release notes
</a>{" "}
to find out what's new!
to find out whats new!
</>
)
}

View File

@@ -8,6 +8,7 @@ import ACLTag from "src/components/acl-tag"
import * as Control from "src/components/control-components"
import NiceIP from "src/components/nice-ip"
import { UpdateAvailableNotification } from "src/components/update-available"
import { AuthResponse, canEdit } from "src/hooks/auth"
import { NodeData } from "src/types"
import Button from "src/ui/button"
import Card from "src/ui/card"
@@ -16,11 +17,11 @@ import QuickCopy from "src/ui/quick-copy"
import { useLocation } from "wouter"
export default function DeviceDetailsView({
readonly,
node,
auth,
}: {
readonly: boolean
node: NodeData
auth: AuthResponse
}) {
return (
<>
@@ -37,11 +38,11 @@ export default function DeviceDetailsView({
})}
/>
</div>
{!readonly && <DisconnectDialog />}
{canEdit("account", auth) && <DisconnectDialog />}
</div>
</Card>
{node.Features["auto-update"] &&
!readonly &&
canEdit("account", auth) &&
node.ClientVersion &&
!node.ClientVersion.RunningLatest && (
<UpdateAvailableNotification details={node.ClientVersion} />

View File

@@ -2,7 +2,7 @@
// SPDX-License-Identifier: BSD-3-Clause
import React from "react"
import { ReactComponent as TailscaleIcon } from "src/assets/icons/tailscale-icon.svg"
import TailscaleIcon from "src/assets/icons/tailscale-icon.svg?react"
/**
* DisconnectedView is rendered after node logout.

View File

@@ -4,21 +4,22 @@
import cx from "classnames"
import React, { useMemo } from "react"
import { apiFetch } from "src/api"
import { ReactComponent as ArrowRight } from "src/assets/icons/arrow-right.svg"
import { ReactComponent as Machine } from "src/assets/icons/machine.svg"
import ArrowRight from "src/assets/icons/arrow-right.svg?react"
import Machine from "src/assets/icons/machine.svg?react"
import AddressCard from "src/components/address-copy-card"
import ExitNodeSelector from "src/components/exit-node-selector"
import { AuthResponse, canEdit } from "src/hooks/auth"
import { NodeData } from "src/types"
import Card from "src/ui/card"
import { pluralize } from "src/utils/util"
import { Link, useLocation } from "wouter"
export default function HomeView({
readonly,
node,
auth,
}: {
readonly: boolean
node: NodeData
auth: AuthResponse
}) {
const [allSubnetRoutes, pendingSubnetRoutes] = useMemo(
() => [
@@ -63,7 +64,11 @@ export default function HomeView({
</div>
{(node.Features["advertise-exit-node"] ||
node.Features["use-exit-node"]) && (
<ExitNodeSelector className="mb-5" node={node} disabled={readonly} />
<ExitNodeSelector
className="mb-5"
node={node}
disabled={!canEdit("exitnodes", auth)}
/>
)}
<Link
className="link font-medium"

View File

@@ -3,7 +3,7 @@
import React, { useState } from "react"
import { useAPI } from "src/api"
import { ReactComponent as TailscaleIcon } from "src/assets/icons/tailscale-icon.svg"
import TailscaleIcon from "src/assets/icons/tailscale-icon.svg?react"
import { NodeData } from "src/types"
import Button from "src/ui/button"
import Collapsible from "src/ui/collapsible"
@@ -41,7 +41,7 @@ export default function LoginView({ data }: { data: NodeData }) {
<>
<div className="mb-6">
<p className="text-gray-700">
Your device's key has expired. Reauthenticate this device by
Your devices key has expired. Reauthenticate this device by
logging in again, or{" "}
<a
href="https://tailscale.com/kb/1028/key-expiry"

View File

@@ -4,9 +4,9 @@
import cx from "classnames"
import React, { useCallback, useMemo, useState } from "react"
import { useAPI } from "src/api"
import { ReactComponent as CheckCircle } from "src/assets/icons/check-circle.svg"
import { ReactComponent as Clock } from "src/assets/icons/clock.svg"
import { ReactComponent as Plus } from "src/assets/icons/plus.svg"
import CheckCircle from "src/assets/icons/check-circle.svg?react"
import Clock from "src/assets/icons/clock.svg?react"
import Plus from "src/assets/icons/plus.svg?react"
import * as Control from "src/components/control-components"
import { NodeData } from "src/types"
import Button from "src/ui/button"

View File

@@ -2,8 +2,8 @@
// SPDX-License-Identifier: BSD-3-Clause
import React from "react"
import { ReactComponent as CheckCircleIcon } from "src/assets/icons/check-circle.svg"
import { ReactComponent as XCircleIcon } from "src/assets/icons/x-circle.svg"
import CheckCircleIcon from "src/assets/icons/check-circle.svg?react"
import XCircleIcon from "src/assets/icons/x-circle.svg?react"
import { ChangelogText } from "src/components/update-available"
import { UpdateState, useInstallUpdate } from "src/hooks/self-update"
import { VersionInfo } from "src/types"
@@ -35,7 +35,7 @@ export function UpdatingView({
<Spinner size="sm" className="text-gray-400" />
<h1 className="text-2xl m-3">Update in progress</h1>
<p className="text-gray-400">
The update shouldn't take more than a couple of minutes. Once it's
The update shouldnt take more than a couple of minutes. Once its
completed, you will be asked to log in again.
</p>
</>

View File

@@ -4,25 +4,50 @@
import { useCallback, useEffect, useState } from "react"
import { apiFetch, setSynoToken } from "src/api"
export enum AuthType {
synology = "synology",
tailscale = "tailscale",
}
export type AuthResponse = {
authNeeded?: AuthType
canManageNode: boolean
serverMode: "login" | "manage"
serverMode: AuthServerMode
authorized: boolean
viewerIdentity?: {
loginName: string
nodeName: string
nodeIP: string
profilePicUrl?: string
capabilities: { [key in PeerCapability]: boolean }
}
needsSynoAuth?: boolean
}
// useAuth reports and refreshes Tailscale auth status
// for the web client.
export type AuthServerMode = "login" | "readonly" | "manage"
export type PeerCapability = "*" | "ssh" | "subnets" | "exitnodes" | "account"
/**
* canEdit reports whether the given auth response specifies that the viewer
* has the ability to edit the given capability.
*/
export function canEdit(cap: PeerCapability, auth: AuthResponse): boolean {
if (!auth.authorized || !auth.viewerIdentity) {
return false
}
if (auth.viewerIdentity.capabilities["*"] === true) {
return true // can edit all features
}
return auth.viewerIdentity.capabilities[cap] === true
}
/**
* hasAnyEditCapabilities reports whether the given auth response specifies
* that the viewer has at least one edit capability. If this is true, the
* user is able to go through the auth flow to authenticate a management
* session.
*/
export function hasAnyEditCapabilities(auth: AuthResponse): boolean {
return Object.values(auth.viewerIdentity?.capabilities || {}).includes(true)
}
/**
* useAuth reports and refreshes Tailscale auth status for the web client.
*/
export default function useAuth() {
const [data, setData] = useState<AuthResponse>()
const [loading, setLoading] = useState<boolean>(true)
@@ -33,18 +58,16 @@ export default function useAuth() {
return apiFetch<AuthResponse>("/auth", "GET")
.then((d) => {
setData(d)
switch (d.authNeeded) {
case AuthType.synology:
fetch("/webman/login.cgi")
.then((r) => r.json())
.then((a) => {
setSynoToken(a.SynoToken)
setRanSynoAuth(true)
setLoading(false)
})
break
default:
setLoading(false)
if (d.needsSynoAuth) {
fetch("/webman/login.cgi")
.then((r) => r.json())
.then((a) => {
setSynoToken(a.SynoToken)
setRanSynoAuth(true)
setLoading(false)
})
} else {
setLoading(false)
}
return d
})
@@ -72,8 +95,13 @@ export default function useAuth() {
useEffect(() => {
loadAuth().then((d) => {
if (!d) {
return
}
if (
!d?.canManageNode &&
!d.authorized &&
hasAnyEditCapabilities(d) &&
// Start auth flow immediately if browser has requested it.
new URLSearchParams(window.location.search).get("check") === "now"
) {
newSession()

View File

@@ -0,0 +1,46 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
import { useCallback, useEffect, useState } from "react"
import { isHTTPS } from "src/utils/util"
import { AuthServerMode } from "./auth"
/**
* useTSWebConnected hook is used to check whether the browser is able to
* connect to the web client served at http://${nodeIPv4}:5252
*/
export function useTSWebConnected(mode: AuthServerMode, nodeIPv4: string) {
const [tsWebConnected, setTSWebConnected] = useState<boolean>(
mode === "manage" // browser already on the web client
)
const [isLoading, setIsLoading] = useState<boolean>(false)
const checkTSWebConnection = useCallback(() => {
if (mode === "manage") {
// Already connected to the web client.
setTSWebConnected(true)
return
}
if (isHTTPS()) {
// When page is loaded over HTTPS, the connectivity check will always
// fail with a mixed-content error. In this case don't bother doing
// the check.
return
}
if (isLoading) {
return // already checking
}
setIsLoading(true)
fetch(`http://${nodeIPv4}:5252/ok`, { mode: "no-cors" })
.then(() => {
setTSWebConnected(true)
setIsLoading(false)
})
.catch(() => setIsLoading(false))
}, [isLoading, mode, nodeIPv4])
// eslint-disable-next-line react-hooks/exhaustive-deps
useEffect(() => checkTSWebConnection(), []) // checking connection for first time on page load
return { tsWebConnected, checkTSWebConnection, isLoading }
}

View File

@@ -3,7 +3,7 @@
import * as Primitive from "@radix-ui/react-collapsible"
import React, { useState } from "react"
import { ReactComponent as ChevronDown } from "src/assets/icons/chevron-down.svg"
import ChevronDown from "src/assets/icons/chevron-down.svg?react"
type CollapsibleProps = {
trigger?: string

View File

@@ -4,7 +4,7 @@
import * as DialogPrimitive from "@radix-ui/react-dialog"
import cx from "classnames"
import React, { Component, ComponentProps, FormEvent } from "react"
import { ReactComponent as X } from "src/assets/icons/x.svg"
import X from "src/assets/icons/x.svg?react"
import Button from "src/ui/button"
import PortalContainerContext from "src/ui/portal-container-context"
import { isObject } from "src/utils/util"

View File

@@ -3,7 +3,7 @@
import cx from "classnames"
import React, { forwardRef, InputHTMLAttributes } from "react"
import { ReactComponent as Search } from "src/assets/icons/search.svg"
import Search from "src/assets/icons/search.svg?react"
type Props = {
className?: string

View File

@@ -10,7 +10,7 @@ import React, {
useState,
} from "react"
import { createPortal } from "react-dom"
import { ReactComponent as X } from "src/assets/icons/x.svg"
import X from "src/assets/icons/x.svg?react"
import { noop } from "src/utils/util"
import { create } from "zustand"
import { shallow } from "zustand/shallow"

View File

@@ -49,3 +49,10 @@ export function isPromise<T = unknown>(val: unknown): val is Promise<T> {
}
return typeof val === "object" && "then" in val
}
/**
* isHTTPS reports whether the current page is loaded over HTTPS.
*/
export function isHTTPS() {
return window.location.protocol === "https:"
}

View File

@@ -1,7 +1,7 @@
const plugin = require("tailwindcss/plugin")
const styles = require("./styles.json")
import plugin from "tailwindcss/plugin"
import styles from "./styles.json"
module.exports = {
const config = {
theme: {
screens: {
sm: "420px",
@@ -96,20 +96,22 @@ module.exports = {
plugins: [
plugin(function ({ addVariant }) {
addVariant("state-open", [
'&[data-state="open"]',
'[data-state="open"] &',
"&[data-state=open”]",
"[data-state=open] &",
])
addVariant("state-closed", [
'&[data-state="closed"]',
'[data-state="closed"] &',
"&[data-state=closed”]",
"[data-state=closed] &",
])
addVariant("state-delayed-open", [
'&[data-state="delayed-open"]',
'[data-state="delayed-open"] &',
"&[data-state=delayed-open”]",
"[data-state=delayed-open] &",
])
addVariant("state-active", ['&[data-state="active"]'])
addVariant("state-inactive", ['&[data-state="inactive"]'])
addVariant("state-active", ["&[data-state=active”]"])
addVariant("state-inactive", ["&[data-state=inactive”]"])
}),
],
content: ["./src/**/*.html", "./src/**/*.{ts,tsx}", "./index.html"],
}
export default config

View File

@@ -5,6 +5,7 @@
"module": "ES2020",
"strict": true,
"sourceMap": true,
"skipLibCheck": true,
"isolatedModules": true,
"moduleResolution": "node",
"forceConsistentCasingInFileNames": true,

View File

@@ -1,6 +1,5 @@
/// <reference types="vitest" />
import { createLogger, defineConfig } from "vite"
import rewrite from "vite-plugin-rewrite-all"
import svgr from "vite-plugin-svgr"
import paths from "vite-tsconfig-paths"
@@ -24,11 +23,6 @@ export default defineConfig({
plugins: [
paths(),
svgr(),
// By default, the Vite dev server doesn't handle dots
// in path names and treats them as static files.
// This plugin changes Vite's routing logic to fix this.
// See: https://github.com/vitejs/vite/issues/2415
rewrite(),
],
build: {
outDir: "build",

View File

@@ -95,6 +95,14 @@ const (
// In this mode, API calls are authenticated via platform auth.
LoginServerMode ServerMode = "login"
// ReadOnlyServerMode is identical to LoginServerMode,
// but does not present a login button to switch to manage mode,
// even if the management client is running and reachable.
//
// This is designed for platforms where the device is configured by other means,
// such as Home Assistant's declarative YAML configuration.
ReadOnlyServerMode ServerMode = "readonly"
// ManageServerMode serves a management client for editing tailscale
// settings of a node.
//
@@ -154,7 +162,7 @@ type ServerOpts struct {
// and not the lifespan of the web server.
func NewServer(opts ServerOpts) (s *Server, err error) {
switch opts.Mode {
case LoginServerMode, ManageServerMode:
case LoginServerMode, ReadOnlyServerMode, ManageServerMode:
// valid types
case "":
return nil, fmt.Errorf("must specify a Mode")
@@ -207,10 +215,14 @@ func NewServer(opts ServerOpts) (s *Server, err error) {
// The client is secured by limiting the interface it listens on,
// or by authenticating requests before they reach the web client.
csrfProtect := csrf.Protect(s.csrfKey(), csrf.Secure(false))
if s.mode == LoginServerMode {
switch s.mode {
case LoginServerMode:
s.apiHandler = csrfProtect(http.HandlerFunc(s.serveLoginAPI))
metric = "web_login_client_initialization"
} else {
case ReadOnlyServerMode:
s.apiHandler = csrfProtect(http.HandlerFunc(s.serveLoginAPI))
metric = "web_readonly_client_initialization"
case ManageServerMode:
s.apiHandler = csrfProtect(http.HandlerFunc(s.serveAPI))
metric = "web_client_initialization"
}
@@ -433,27 +445,198 @@ func (s *Server) serveLoginAPI(w http.ResponseWriter, r *http.Request) {
}
}
type authType string
type apiHandler[data any] struct {
s *Server
w http.ResponseWriter
r *http.Request
var (
synoAuth authType = "synology" // user needs a SynoToken for subsequent API calls
tailscaleAuth authType = "tailscale" // user needs to complete Tailscale check mode
)
// permissionCheck allows for defining whether a requesting peer's
// capabilities grant them access to make the given data update.
// If permissionCheck reports false, the request fails as unauthorized.
permissionCheck func(data data, peer peerCapabilities) bool
}
// newHandler constructs a new api handler which restricts the given request
// to the specified permission check. If the permission check fails for
// the peer associated with the request, an unauthorized error is returned
// to the client.
func newHandler[data any](s *Server, w http.ResponseWriter, r *http.Request, permissionCheck func(data data, peer peerCapabilities) bool) *apiHandler[data] {
return &apiHandler[data]{
s: s,
w: w,
r: r,
permissionCheck: permissionCheck,
}
}
// alwaysAllowed can be passed as the permissionCheck argument to newHandler
// for requests that are always allowed to complete regardless of a peer's
// capabilities.
func alwaysAllowed[data any](_ data, _ peerCapabilities) bool { return true }
func (a *apiHandler[data]) getPeer() (peerCapabilities, error) {
// TODO(tailscale/corp#16695,sonia): We also call StatusWithoutPeers and
// WhoIs when originally checking for a session from authorizeRequest.
// Would be nice if we could pipe those through to here so we don't end
// up having to re-call them to grab the peer capabilities.
status, err := a.s.lc.StatusWithoutPeers(a.r.Context())
if err != nil {
return nil, err
}
whois, err := a.s.lc.WhoIs(a.r.Context(), a.r.RemoteAddr)
if err != nil {
return nil, err
}
peer, err := toPeerCapabilities(status, whois)
if err != nil {
return nil, err
}
return peer, nil
}
type noBodyData any // empty type, for use from serveAPI for endpoints with empty body
// handle runs the given handler if the source peer satisfies the
// constraints for running this request.
//
// handle is expected for use when `data` type is empty, or set to
// `noBodyData` in practice. For requests that expect JSON body data
// to be attached, use handleJSON instead.
func (a *apiHandler[data]) handle(h http.HandlerFunc) {
peer, err := a.getPeer()
if err != nil {
http.Error(a.w, err.Error(), http.StatusInternalServerError)
return
}
var body data // not used
if !a.permissionCheck(body, peer) {
http.Error(a.w, "not allowed", http.StatusUnauthorized)
return
}
h(a.w, a.r)
}
// handleJSON manages decoding the request's body JSON and passing
// it on to the provided function if the source peer satisfies the
// constraints for running this request.
func (a *apiHandler[data]) handleJSON(h func(ctx context.Context, data data) error) {
defer a.r.Body.Close()
var body data
if err := json.NewDecoder(a.r.Body).Decode(&body); err != nil {
http.Error(a.w, err.Error(), http.StatusInternalServerError)
return
}
peer, err := a.getPeer()
if err != nil {
http.Error(a.w, err.Error(), http.StatusInternalServerError)
return
}
if !a.permissionCheck(body, peer) {
http.Error(a.w, "not allowed", http.StatusUnauthorized)
return
}
if err := h(a.r.Context(), body); err != nil {
http.Error(a.w, err.Error(), http.StatusInternalServerError)
return
}
a.w.WriteHeader(http.StatusOK)
}
// serveAPI serves requests for the web client api.
// It should only be called by Server.ServeHTTP, via Server.apiHandler,
// which protects the handler using gorilla csrf.
func (s *Server) serveAPI(w http.ResponseWriter, r *http.Request) {
if r.Method == httpm.PATCH {
// Enforce that PATCH requests are always application/json.
if ct := r.Header.Get("Content-Type"); ct != "application/json" {
http.Error(w, "invalid request", http.StatusBadRequest)
return
}
}
w.Header().Set("X-CSRF-Token", csrf.Token(r))
path := strings.TrimPrefix(r.URL.Path, "/api")
switch {
case path == "/data" && r.Method == httpm.GET:
newHandler[noBodyData](s, w, r, alwaysAllowed).
handle(s.serveGetNodeData)
return
case path == "/exit-nodes" && r.Method == httpm.GET:
newHandler[noBodyData](s, w, r, alwaysAllowed).
handle(s.serveGetExitNodes)
return
case path == "/routes" && r.Method == httpm.POST:
peerAllowed := func(d postRoutesRequest, p peerCapabilities) bool {
if d.SetExitNode && !p.canEdit(capFeatureExitNodes) {
return false
} else if d.SetRoutes && !p.canEdit(capFeatureSubnets) {
return false
}
return true
}
newHandler[postRoutesRequest](s, w, r, peerAllowed).
handleJSON(s.servePostRoutes)
return
case path == "/device-details-click" && r.Method == httpm.POST:
newHandler[noBodyData](s, w, r, alwaysAllowed).
handle(s.serveDeviceDetailsClick)
return
case path == "/local/v0/logout" && r.Method == httpm.POST:
peerAllowed := func(_ noBodyData, peer peerCapabilities) bool {
return peer.canEdit(capFeatureAccount)
}
newHandler[noBodyData](s, w, r, peerAllowed).
handle(s.proxyRequestToLocalAPI)
return
case path == "/local/v0/prefs" && r.Method == httpm.PATCH:
peerAllowed := func(data maskedPrefs, peer peerCapabilities) bool {
if data.RunSSHSet && !peer.canEdit(capFeatureSSH) {
return false
}
return true
}
newHandler[maskedPrefs](s, w, r, peerAllowed).
handleJSON(s.serveUpdatePrefs)
return
case path == "/local/v0/update/check" && r.Method == httpm.GET:
newHandler[noBodyData](s, w, r, alwaysAllowed).
handle(s.proxyRequestToLocalAPI)
return
case path == "/local/v0/update/check" && r.Method == httpm.POST:
peerAllowed := func(_ noBodyData, peer peerCapabilities) bool {
return peer.canEdit(capFeatureAccount)
}
newHandler[noBodyData](s, w, r, peerAllowed).
handle(s.proxyRequestToLocalAPI)
return
case path == "/local/v0/update/progress" && r.Method == httpm.POST:
newHandler[noBodyData](s, w, r, alwaysAllowed).
handle(s.proxyRequestToLocalAPI)
return
case path == "/local/v0/upload-client-metrics" && r.Method == httpm.POST:
newHandler[noBodyData](s, w, r, alwaysAllowed).
handle(s.proxyRequestToLocalAPI)
return
}
http.Error(w, "invalid endpoint", http.StatusNotFound)
}
type authResponse struct {
AuthNeeded authType `json:"authNeeded,omitempty"` // filled when user needs to complete a specific type of auth
CanManageNode bool `json:"canManageNode"`
ViewerIdentity *viewerIdentity `json:"viewerIdentity,omitempty"`
ServerMode ServerMode `json:"serverMode"`
Authorized bool `json:"authorized"` // has an authorized management session
ViewerIdentity *viewerIdentity `json:"viewerIdentity,omitempty"`
NeedsSynoAuth bool `json:"needsSynoAuth,omitempty"`
}
// viewerIdentity is the Tailscale identity of the source node
// connected to this web client.
type viewerIdentity struct {
LoginName string `json:"loginName"`
NodeName string `json:"nodeName"`
NodeIP string `json:"nodeIP"`
ProfilePicURL string `json:"profilePicUrl,omitempty"`
LoginName string `json:"loginName"`
NodeName string `json:"nodeName"`
NodeIP string `json:"nodeIP"`
ProfilePicURL string `json:"profilePicUrl,omitempty"`
Capabilities peerCapabilities `json:"capabilities"` // features peer is allowed to edit
}
// serverAPIAuth handles requests to the /api/auth endpoint
@@ -462,12 +645,20 @@ func (s *Server) serveAPIAuth(w http.ResponseWriter, r *http.Request) {
var resp authResponse
resp.ServerMode = s.mode
session, whois, status, sErr := s.getSession(r)
var caps peerCapabilities
if whois != nil {
var err error
caps, err = toPeerCapabilities(status, whois)
if err != nil {
http.Error(w, sErr.Error(), http.StatusInternalServerError)
return
}
resp.ViewerIdentity = &viewerIdentity{
LoginName: whois.UserProfile.LoginName,
NodeName: whois.Node.Name,
ProfilePicURL: whois.UserProfile.ProfilePicURL,
Capabilities: caps,
}
if addrs := whois.Node.Addresses; len(addrs) > 0 {
resp.ViewerIdentity.NodeIP = addrs[0].Addr().String()
@@ -476,7 +667,7 @@ func (s *Server) serveAPIAuth(w http.ResponseWriter, r *http.Request) {
// First verify platform auth.
// If platform auth is needed, this should happen first.
if s.mode == LoginServerMode {
if s.mode == LoginServerMode || s.mode == ReadOnlyServerMode {
switch distro.Get() {
case distro.Synology:
authorized, err := authorizeSynology(r)
@@ -485,7 +676,7 @@ func (s *Server) serveAPIAuth(w http.ResponseWriter, r *http.Request) {
return
}
if !authorized {
resp.AuthNeeded = synoAuth
resp.NeedsSynoAuth = true
writeJSON(w, resp)
return
}
@@ -501,21 +692,17 @@ func (s *Server) serveAPIAuth(w http.ResponseWriter, r *http.Request) {
switch {
case sErr != nil && errors.Is(sErr, errNotUsingTailscale):
// Restricted to the readonly view, no auth action to take.
s.lc.IncrementCounter(r.Context(), "web_client_viewing_local", 1)
resp.AuthNeeded = ""
resp.Authorized = false // restricted to the readonly view
case sErr != nil && errors.Is(sErr, errNotOwner):
// Restricted to the readonly view, no auth action to take.
s.lc.IncrementCounter(r.Context(), "web_client_viewing_not_owner", 1)
resp.AuthNeeded = ""
resp.Authorized = false // restricted to the readonly view
case sErr != nil && errors.Is(sErr, errTaggedLocalSource):
// Restricted to the readonly view, no auth action to take.
s.lc.IncrementCounter(r.Context(), "web_client_viewing_local_tag", 1)
resp.AuthNeeded = ""
resp.Authorized = false // restricted to the readonly view
case sErr != nil && errors.Is(sErr, errTaggedRemoteSource):
// Restricted to the readonly view, no auth action to take.
s.lc.IncrementCounter(r.Context(), "web_client_viewing_remote_tag", 1)
resp.AuthNeeded = ""
resp.Authorized = false // restricted to the readonly view
case sErr != nil && !errors.Is(sErr, errNoSession):
// Any other error.
http.Error(w, sErr.Error(), http.StatusInternalServerError)
@@ -526,16 +713,26 @@ func (s *Server) serveAPIAuth(w http.ResponseWriter, r *http.Request) {
} else {
s.lc.IncrementCounter(r.Context(), "web_client_managing_remote", 1)
}
resp.CanManageNode = true
resp.AuthNeeded = ""
// User has a valid session. They're now authorized to edit if they
// have any edit capabilities. In practice, they won't be sent through
// the auth flow if they don't have edit caps, but their ACL granted
// permissions may change at any time. The frontend views and backend
// endpoints are always restricted to their current capabilities in
// addition to a valid session.
//
// But, we also check the caps here for a better user experience on
// the frontend login toggle, which uses resp.Authorized to display
// "viewing" vs "managing" copy. If they don't have caps, we want to
// display "viewing" even if they have a valid session.
resp.Authorized = !caps.isEmpty()
default:
// whois being nil implies local as the request did not come over Tailscale
if whois == nil || (whois.Node.StableID == status.Self.ID) {
// whois being nil implies local as the request did not come over Tailscale.
s.lc.IncrementCounter(r.Context(), "web_client_viewing_local", 1)
} else {
s.lc.IncrementCounter(r.Context(), "web_client_viewing_remote", 1)
}
resp.AuthNeeded = tailscaleAuth
resp.Authorized = false // not yet authorized
}
writeJSON(w, resp)
@@ -599,32 +796,6 @@ func (s *Server) serveAPIAuthSessionWait(w http.ResponseWriter, r *http.Request)
}
}
// serveAPI serves requests for the web client api.
// It should only be called by Server.ServeHTTP, via Server.apiHandler,
// which protects the handler using gorilla csrf.
func (s *Server) serveAPI(w http.ResponseWriter, r *http.Request) {
w.Header().Set("X-CSRF-Token", csrf.Token(r))
path := strings.TrimPrefix(r.URL.Path, "/api")
switch {
case path == "/data" && r.Method == httpm.GET:
s.serveGetNodeData(w, r)
return
case path == "/exit-nodes" && r.Method == httpm.GET:
s.serveGetExitNodes(w, r)
return
case path == "/routes" && r.Method == httpm.POST:
s.servePostRoutes(w, r)
return
case path == "/device-details-click" && r.Method == httpm.POST:
s.serveDeviceDetailsClick(w, r)
return
case strings.HasPrefix(path, "/local/"):
s.proxyRequestToLocalAPI(w, r)
return
}
http.Error(w, "invalid endpoint", http.StatusNotFound)
}
type nodeData struct {
ID tailcfg.StableNodeID
Status string
@@ -861,6 +1032,23 @@ func (s *Server) serveGetExitNodes(w http.ResponseWriter, r *http.Request) {
writeJSON(w, exitNodes)
}
// maskedPrefs is the subset of ipn.MaskedPrefs that are
// allowed to be editable via the web UI.
type maskedPrefs struct {
RunSSHSet bool
RunSSH bool
}
func (s *Server) serveUpdatePrefs(ctx context.Context, prefs maskedPrefs) error {
_, err := s.lc.EditPrefs(ctx, &ipn.MaskedPrefs{
RunSSHSet: prefs.RunSSHSet,
Prefs: ipn.Prefs{
RunSSH: prefs.RunSSH,
},
})
return err
}
type postRoutesRequest struct {
SetExitNode bool // when set, UseExitNode and AdvertiseExitNode values are applied
SetRoutes bool // when set, AdvertiseRoutes value is applied
@@ -869,18 +1057,10 @@ type postRoutesRequest struct {
AdvertiseRoutes []string
}
func (s *Server) servePostRoutes(w http.ResponseWriter, r *http.Request) {
defer r.Body.Close()
var data postRoutesRequest
if err := json.NewDecoder(r.Body).Decode(&data); err != nil {
http.Error(w, err.Error(), http.StatusInternalServerError)
return
}
prefs, err := s.lc.GetPrefs(r.Context())
func (s *Server) servePostRoutes(ctx context.Context, data postRoutesRequest) error {
prefs, err := s.lc.GetPrefs(ctx)
if err != nil {
http.Error(w, err.Error(), http.StatusInternalServerError)
return
return err
}
var currNonExitRoutes []string
var currAdvertisingExitNode bool
@@ -903,8 +1083,7 @@ func (s *Server) servePostRoutes(w http.ResponseWriter, r *http.Request) {
routesStr := strings.Join(data.AdvertiseRoutes, ",")
routes, err := netutil.CalcAdvertiseRoutes(routesStr, data.AdvertiseExitNode)
if err != nil {
http.Error(w, err.Error(), http.StatusInternalServerError)
return
return err
}
hasExitNodeRoute := func(all []netip.Prefix) bool {
@@ -913,8 +1092,7 @@ func (s *Server) servePostRoutes(w http.ResponseWriter, r *http.Request) {
}
if !data.UseExitNode.IsZero() && hasExitNodeRoute(routes) {
http.Error(w, "cannot use and advertise exit node at same time", http.StatusBadRequest)
return
return errors.New("cannot use and advertise exit node at same time")
}
// Make prefs update.
@@ -926,12 +1104,8 @@ func (s *Server) servePostRoutes(w http.ResponseWriter, r *http.Request) {
AdvertiseRoutes: routes,
},
}
if _, err := s.lc.EditPrefs(r.Context(), p); err != nil {
http.Error(w, err.Error(), http.StatusInternalServerError)
return
}
w.WriteHeader(http.StatusOK)
_, err = s.lc.EditPrefs(ctx, p)
return err
}
// tailscaleUp starts the daemon with the provided options.
@@ -1070,26 +1244,12 @@ func (s *Server) serveDeviceDetailsClick(w http.ResponseWriter, r *http.Request)
//
// The web API request path is expected to exactly match a localapi path,
// with prefix /api/local/ rather than /localapi/.
//
// If the localapi path is not included in localapiAllowlist,
// the request is rejected.
func (s *Server) proxyRequestToLocalAPI(w http.ResponseWriter, r *http.Request) {
path := strings.TrimPrefix(r.URL.Path, "/api/local")
if r.URL.Path == path { // missing prefix
http.Error(w, "invalid request", http.StatusBadRequest)
return
}
if r.Method == httpm.PATCH {
// enforce that PATCH requests are always application/json
if ct := r.Header.Get("Content-Type"); ct != "application/json" {
http.Error(w, "invalid request", http.StatusBadRequest)
return
}
}
if !slices.Contains(localapiAllowlist, path) {
http.Error(w, fmt.Sprintf("%s not allowed from localapi proxy", path), http.StatusForbidden)
return
}
localAPIURL := "http://" + apitype.LocalAPIHost + "/localapi" + path
req, err := http.NewRequestWithContext(r.Context(), r.Method, localAPIURL, r.Body)
@@ -1114,21 +1274,6 @@ func (s *Server) proxyRequestToLocalAPI(w http.ResponseWriter, r *http.Request)
}
}
// localapiAllowlist is an allowlist of localapi endpoints the
// web client is allowed to proxy to the client's localapi.
//
// Rather than exposing all localapi endpoints over the proxy,
// this limits to just the ones actually used from the web
// client frontend.
var localapiAllowlist = []string{
"/v0/logout",
"/v0/prefs",
"/v0/update/check",
"/v0/update/install",
"/v0/update/progress",
"/v0/upload-client-metrics",
}
// csrfKey returns a key that can be used for CSRF protection.
// If an error occurs during key creation, the error is logged and the active process terminated.
// If the server is running in CGI mode, the key is cached to disk and reused between requests.

View File

@@ -4,6 +4,7 @@
package web
import (
"bytes"
"context"
"encoding/json"
"errors"
@@ -86,75 +87,172 @@ func TestQnapAuthnURL(t *testing.T) {
// TestServeAPI tests the web client api's handling of
// 1. invalid endpoint errors
// 2. localapi proxy allowlist
// 2. permissioning of api endpoints based on node capabilities
func TestServeAPI(t *testing.T) {
selfTags := views.SliceOf([]string{"tag:server"})
self := &ipnstate.PeerStatus{ID: "self", Tags: &selfTags}
prefs := &ipn.Prefs{}
remoteUser := &tailcfg.UserProfile{ID: tailcfg.UserID(1)}
remoteIPWithAllCapabilities := "100.100.100.101"
remoteIPWithNoCapabilities := "100.100.100.102"
lal := memnet.Listen("local-tailscaled.sock:80")
defer lal.Close()
// Serve dummy localapi. Just returns "success".
localapi := &http.Server{Handler: http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
fmt.Fprintf(w, "success")
})}
localapi := mockLocalAPI(t,
map[string]*apitype.WhoIsResponse{
remoteIPWithAllCapabilities: {
Node: &tailcfg.Node{StableID: "node1"},
UserProfile: remoteUser,
CapMap: tailcfg.PeerCapMap{tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{"{\"canEdit\":[\"*\"]}"}},
},
remoteIPWithNoCapabilities: {
Node: &tailcfg.Node{StableID: "node2"},
UserProfile: remoteUser,
},
},
func() *ipnstate.PeerStatus { return self },
func() *ipn.Prefs { return prefs },
nil,
)
defer localapi.Close()
go localapi.Serve(lal)
s := &Server{lc: &tailscale.LocalClient{Dial: lal.Dial}}
s := &Server{
mode: ManageServerMode,
lc: &tailscale.LocalClient{Dial: lal.Dial},
timeNow: time.Now,
}
type requestTest struct {
remoteIP string
wantResponse string
wantStatus int
}
tests := []struct {
name string
reqMethod string
reqPath string
reqMethod string
reqContentType string
wantResp string
wantStatus int
reqBody string
tests []requestTest
}{{
name: "invalid_endpoint",
reqMethod: httpm.POST,
reqPath: "/not-an-endpoint",
wantResp: "invalid endpoint",
wantStatus: http.StatusNotFound,
reqPath: "/not-an-endpoint",
reqMethod: httpm.POST,
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "invalid endpoint",
wantStatus: http.StatusNotFound,
}, {
remoteIP: remoteIPWithAllCapabilities,
wantResponse: "invalid endpoint",
wantStatus: http.StatusNotFound,
}},
}, {
name: "not_in_localapi_allowlist",
reqMethod: httpm.POST,
reqPath: "/local/v0/not-allowlisted",
wantResp: "/v0/not-allowlisted not allowed from localapi proxy",
wantStatus: http.StatusForbidden,
reqPath: "/local/v0/not-an-endpoint",
reqMethod: httpm.POST,
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "invalid endpoint",
wantStatus: http.StatusNotFound,
}, {
remoteIP: remoteIPWithAllCapabilities,
wantResponse: "invalid endpoint",
wantStatus: http.StatusNotFound,
}},
}, {
name: "in_localapi_allowlist",
reqMethod: httpm.POST,
reqPath: "/local/v0/logout",
wantResp: "success", // Successfully allowed to hit localapi.
wantStatus: http.StatusOK,
reqPath: "/local/v0/logout",
reqMethod: httpm.POST,
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "not allowed", // requesting node has insufficient permissions
wantStatus: http.StatusUnauthorized,
}, {
remoteIP: remoteIPWithAllCapabilities,
wantResponse: "success", // requesting node has sufficient permissions
wantStatus: http.StatusOK,
}},
}, {
reqPath: "/exit-nodes",
reqMethod: httpm.GET,
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "null",
wantStatus: http.StatusOK, // allowed, no additional capabilities required
}, {
remoteIP: remoteIPWithAllCapabilities,
wantResponse: "null",
wantStatus: http.StatusOK,
}},
}, {
reqPath: "/routes",
reqMethod: httpm.POST,
reqBody: "{\"setExitNode\":true}",
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "not allowed",
wantStatus: http.StatusUnauthorized,
}, {
remoteIP: remoteIPWithAllCapabilities,
wantStatus: http.StatusOK,
}},
}, {
name: "patch_bad_contenttype",
reqMethod: httpm.PATCH,
reqPath: "/local/v0/prefs",
reqMethod: httpm.PATCH,
reqBody: "{\"runSSHSet\":true}",
reqContentType: "application/json",
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "not allowed",
wantStatus: http.StatusUnauthorized,
}, {
remoteIP: remoteIPWithAllCapabilities,
wantStatus: http.StatusOK,
}},
}, {
reqPath: "/local/v0/prefs",
reqMethod: httpm.PATCH,
reqContentType: "multipart/form-data",
wantResp: "invalid request",
wantStatus: http.StatusBadRequest,
tests: []requestTest{{
remoteIP: remoteIPWithNoCapabilities,
wantResponse: "invalid request",
wantStatus: http.StatusBadRequest,
}, {
remoteIP: remoteIPWithAllCapabilities,
wantResponse: "invalid request",
wantStatus: http.StatusBadRequest,
}},
}}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
r := httptest.NewRequest(tt.reqMethod, "/api"+tt.reqPath, nil)
if tt.reqContentType != "" {
r.Header.Add("Content-Type", tt.reqContentType)
}
w := httptest.NewRecorder()
for _, req := range tt.tests {
t.Run(req.remoteIP+"_requesting_"+tt.reqPath, func(t *testing.T) {
var reqBody io.Reader
if tt.reqBody != "" {
reqBody = bytes.NewBuffer([]byte(tt.reqBody))
}
r := httptest.NewRequest(tt.reqMethod, "/api"+tt.reqPath, reqBody)
r.RemoteAddr = req.remoteIP
if tt.reqContentType != "" {
r.Header.Add("Content-Type", tt.reqContentType)
}
w := httptest.NewRecorder()
s.serveAPI(w, r)
res := w.Result()
defer res.Body.Close()
if gotStatus := res.StatusCode; tt.wantStatus != gotStatus {
t.Errorf("wrong status; want=%v, got=%v", tt.wantStatus, gotStatus)
}
body, err := io.ReadAll(res.Body)
if err != nil {
t.Fatal(err)
}
gotResp := strings.TrimSuffix(string(body), "\n") // trim trailing newline
if tt.wantResp != gotResp {
t.Errorf("wrong response; want=%q, got=%q", tt.wantResp, gotResp)
}
})
s.serveAPI(w, r)
res := w.Result()
defer res.Body.Close()
if gotStatus := res.StatusCode; req.wantStatus != gotStatus {
t.Errorf("wrong status; want=%v, got=%v", req.wantStatus, gotStatus)
}
body, err := io.ReadAll(res.Body)
if err != nil {
t.Fatal(err)
}
gotResp := strings.TrimSuffix(string(body), "\n") // trim trailing newline
if req.wantResponse != gotResp {
t.Errorf("wrong response; want=%q, got=%q", req.wantResponse, gotResp)
}
})
}
}
}
@@ -450,6 +548,7 @@ func TestServeAuth(t *testing.T) {
NodeName: remoteNode.Node.Name,
NodeIP: remoteIP,
ProfilePicURL: user.ProfilePicURL,
Capabilities: peerCapabilities{capFeatureAll: true},
}
testControlURL := &defaultControlURL
@@ -523,7 +622,7 @@ func TestServeAuth(t *testing.T) {
name: "no-session",
path: "/api/auth",
wantStatus: http.StatusOK,
wantResp: &authResponse{AuthNeeded: tailscaleAuth, ViewerIdentity: vi, ServerMode: ManageServerMode},
wantResp: &authResponse{ViewerIdentity: vi, ServerMode: ManageServerMode},
wantNewCookie: false,
wantSession: nil,
},
@@ -548,7 +647,7 @@ func TestServeAuth(t *testing.T) {
path: "/api/auth",
cookie: successCookie,
wantStatus: http.StatusOK,
wantResp: &authResponse{AuthNeeded: tailscaleAuth, ViewerIdentity: vi, ServerMode: ManageServerMode},
wantResp: &authResponse{ViewerIdentity: vi, ServerMode: ManageServerMode},
wantSession: &browserSession{
ID: successCookie,
SrcNode: remoteNode.Node.ID,
@@ -596,7 +695,7 @@ func TestServeAuth(t *testing.T) {
path: "/api/auth",
cookie: successCookie,
wantStatus: http.StatusOK,
wantResp: &authResponse{CanManageNode: true, ViewerIdentity: vi, ServerMode: ManageServerMode},
wantResp: &authResponse{Authorized: true, ViewerIdentity: vi, ServerMode: ManageServerMode},
wantSession: &browserSession{
ID: successCookie,
SrcNode: remoteNode.Node.ID,
@@ -1097,6 +1196,217 @@ func TestRequireTailscaleIP(t *testing.T) {
}
}
func TestPeerCapabilities(t *testing.T) {
userOwnedStatus := &ipnstate.Status{Self: &ipnstate.PeerStatus{UserID: tailcfg.UserID(1)}}
tags := views.SliceOf[string]([]string{"tag:server"})
tagOwnedStatus := &ipnstate.Status{Self: &ipnstate.PeerStatus{Tags: &tags}}
// Testing web.toPeerCapabilities
toPeerCapsTests := []struct {
name string
status *ipnstate.Status
whois *apitype.WhoIsResponse
wantCaps peerCapabilities
}{
{
name: "empty-whois",
status: userOwnedStatus,
whois: nil,
wantCaps: peerCapabilities{},
},
{
name: "user-owned-node-non-owner-caps-ignored",
status: userOwnedStatus,
whois: &apitype.WhoIsResponse{
UserProfile: &tailcfg.UserProfile{ID: tailcfg.UserID(2)},
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"ssh\",\"subnets\"]}",
},
},
},
wantCaps: peerCapabilities{},
},
{
name: "user-owned-node-owner-caps-ignored",
status: userOwnedStatus,
whois: &apitype.WhoIsResponse{
UserProfile: &tailcfg.UserProfile{ID: tailcfg.UserID(1)},
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"ssh\",\"subnets\"]}",
},
},
},
wantCaps: peerCapabilities{capFeatureAll: true}, // should just have wildcard
},
{
name: "tag-owned-no-webui-caps",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityDebugPeer: []tailcfg.RawMessage{},
},
},
wantCaps: peerCapabilities{},
},
{
name: "tag-owned-one-webui-cap",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"ssh\",\"subnets\"]}",
},
},
},
wantCaps: peerCapabilities{
capFeatureSSH: true,
capFeatureSubnets: true,
},
},
{
name: "tag-owned-multiple-webui-cap",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"ssh\",\"subnets\"]}",
"{\"canEdit\":[\"subnets\",\"exitnodes\",\"*\"]}",
},
},
},
wantCaps: peerCapabilities{
capFeatureSSH: true,
capFeatureSubnets: true,
capFeatureExitNodes: true,
capFeatureAll: true,
},
},
{
name: "tag-owned-case-insensitive-caps",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"SSH\",\"sUBnets\"]}",
},
},
},
wantCaps: peerCapabilities{
capFeatureSSH: true,
capFeatureSubnets: true,
},
},
{
name: "tag-owned-random-canEdit-contents-get-dropped",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"unknown-feature\"]}",
},
},
},
wantCaps: peerCapabilities{},
},
{
name: "tag-owned-no-canEdit-section",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1)},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canDoSomething\":[\"*\"]}",
},
},
},
wantCaps: peerCapabilities{},
},
{
name: "tagged-source-caps-ignored",
status: tagOwnedStatus,
whois: &apitype.WhoIsResponse{
Node: &tailcfg.Node{ID: tailcfg.NodeID(1), Tags: tags.AsSlice()},
CapMap: tailcfg.PeerCapMap{
tailcfg.PeerCapabilityWebUI: []tailcfg.RawMessage{
"{\"canEdit\":[\"ssh\",\"subnets\"]}",
},
},
},
wantCaps: peerCapabilities{},
},
}
for _, tt := range toPeerCapsTests {
t.Run("toPeerCapabilities-"+tt.name, func(t *testing.T) {
got, err := toPeerCapabilities(tt.status, tt.whois)
if err != nil {
t.Fatalf("unexpected: %v", err)
}
if diff := cmp.Diff(got, tt.wantCaps); diff != "" {
t.Errorf("wrong caps; (-got+want):%v", diff)
}
})
}
// Testing web.peerCapabilities.canEdit
canEditTests := []struct {
name string
caps peerCapabilities
wantCanEdit map[capFeature]bool
}{
{
name: "empty-caps",
caps: nil,
wantCanEdit: map[capFeature]bool{
capFeatureAll: false,
capFeatureSSH: false,
capFeatureSubnets: false,
capFeatureExitNodes: false,
capFeatureAccount: false,
},
},
{
name: "some-caps",
caps: peerCapabilities{capFeatureSSH: true, capFeatureAccount: true},
wantCanEdit: map[capFeature]bool{
capFeatureAll: false,
capFeatureSSH: true,
capFeatureSubnets: false,
capFeatureExitNodes: false,
capFeatureAccount: true,
},
},
{
name: "wildcard-in-caps",
caps: peerCapabilities{capFeatureAll: true, capFeatureAccount: true},
wantCanEdit: map[capFeature]bool{
capFeatureAll: true,
capFeatureSSH: true,
capFeatureSubnets: true,
capFeatureExitNodes: true,
capFeatureAccount: true,
},
},
}
for _, tt := range canEditTests {
t.Run("canEdit-"+tt.name, func(t *testing.T) {
for f, want := range tt.wantCanEdit {
if got := tt.caps.canEdit(f); got != want {
t.Errorf("wrong canEdit(%s); got=%v, want=%v", f, got, want)
}
}
})
}
}
var (
defaultControlURL = "https://controlplane.tailscale.com"
testAuthPath = "/a/12345"
@@ -1143,6 +1453,9 @@ func mockLocalAPI(t *testing.T, whoIs map[string]*apitype.WhoIsResponse, self fu
metricCapture(metricNames[0].Name)
writeJSON(w, struct{}{})
return
case "/localapi/v0/logout":
fmt.Fprintf(w, "success")
return
default:
t.Fatalf("unhandled localapi test endpoint %q, add to localapi handler func in test", r.URL.Path)
}

File diff suppressed because it is too large Load Diff

View File

@@ -167,6 +167,10 @@ func (up *Updater) getUpdateFunction() (fn updateFunction, canAutoUpdate bool) {
return up.updateWindows, true
case "linux":
switch distro.Get() {
case distro.NixOS:
// NixOS packages are immutable and managed with a system-wide
// configuration.
return up.updateNixos, false
case distro.Synology:
// Synology updates use our own pkgs.tailscale.com instead of the
// Synology Package Center. We should eventually get to a regular
@@ -522,6 +526,13 @@ func (up *Updater) updateArchLike() error {
you can use "pacman --sync --refresh --sysupgrade" or "pacman -Syu" to upgrade the system, including Tailscale.`)
}
func (up *Updater) updateNixos() error {
// NixOS package updates are managed on a system level and not individually.
// Direct users to update their nix channel or nixpkgs flake input to
// receive the latest version.
return errors.New(`individual package updates are not supported on NixOS installations. Update your system channel or flake inputs to get the latest Tailscale version from nixpkgs.`)
}
const yumRepoConfigFile = "/etc/yum.repos.d/tailscale.repo"
// updateFedoraLike updates tailscale on any distros in the Fedora family,
@@ -654,6 +665,7 @@ func (up *Updater) updateAlpineLike() (err error) {
func parseAlpinePackageVersion(out []byte) (string, error) {
s := bufio.NewScanner(bytes.NewReader(out))
var maxVer string
for s.Scan() {
// The line should look like this:
// tailscale-1.44.2-r0 description:
@@ -665,7 +677,13 @@ func parseAlpinePackageVersion(out []byte) (string, error) {
if len(parts) < 3 {
return "", fmt.Errorf("malformed info line: %q", line)
}
return parts[1], nil
ver := parts[1]
if cmpver.Compare(ver, maxVer) == 1 {
maxVer = ver
}
}
if maxVer != "" {
return maxVer, nil
}
return "", errors.New("tailscale version not found in output")
}
@@ -818,7 +836,7 @@ func (up *Updater) switchOutputToFile() (io.Closer, error) {
func (up *Updater) installMSI(msi string) error {
var err error
for tries := 0; tries < 2; tries++ {
cmd := exec.Command("msiexec.exe", "/i", filepath.Base(msi), "/quiet", "/promptrestart", "/qn")
cmd := exec.Command("msiexec.exe", "/i", filepath.Base(msi), "/quiet", "/norestart", "/qn")
cmd.Dir = filepath.Dir(msi)
cmd.Stdout = up.Stdout
cmd.Stderr = up.Stderr

View File

@@ -251,6 +251,29 @@ tailscale installed size:
out: "",
wantErr: true,
},
{
desc: "multiple versions",
out: `
tailscale-1.54.1-r0 description:
The easiest, most secure way to use WireGuard and 2FA
tailscale-1.54.1-r0 webpage:
https://tailscale.com/
tailscale-1.54.1-r0 installed size:
34 MiB
tailscale-1.58.2-r0 description:
The easiest, most secure way to use WireGuard and 2FA
tailscale-1.58.2-r0 webpage:
https://tailscale.com/
tailscale-1.58.2-r0 installed size:
35 MiB
`,
want: "1.58.2",
},
}
for _, tt := range tests {

View File

@@ -55,6 +55,15 @@
// and not `tailscale up` or `tailscale set`.
// The config file contents are currently read once on container start.
// NB: This env var is currently experimental and the logic will likely change!
// - EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS: if set to true
// and if this containerboot instance is an L7 ingress proxy (created by
// the Kubernetes operator), set up rules to allow proxying cluster traffic,
// received on the Pod IP of this node, to the ingress target in the cluster.
// This, in conjunction with MagicDNS name resolution in cluster, can be
// useful for cases where a cluster workload needs to access a target in
// cluster using the same hostname (in this case, the MagicDNS name of the ingress proxy)
// as a non-cluster workload on tailnet.
// This is only meant to be configured by the Kubernetes operator.
//
// When running on Kubernetes, containerboot defaults to storing state in the
// "tailscale" kube secret. To store state on local disk instead, set
@@ -108,29 +117,31 @@ func newNetfilterRunner(logf logger.Logf) (linuxfw.NetfilterRunner, error) {
func main() {
log.SetPrefix("boot: ")
tailscale.I_Acknowledge_This_API_Is_Unstable = true
cfg := &settings{
AuthKey: defaultEnvs([]string{"TS_AUTHKEY", "TS_AUTH_KEY"}, ""),
Hostname: defaultEnv("TS_HOSTNAME", ""),
Routes: defaultEnvStringPointer("TS_ROUTES"),
ServeConfigPath: defaultEnv("TS_SERVE_CONFIG", ""),
ProxyTo: defaultEnv("TS_DEST_IP", ""),
TailnetTargetIP: defaultEnv("TS_TAILNET_TARGET_IP", ""),
TailnetTargetFQDN: defaultEnv("TS_TAILNET_TARGET_FQDN", ""),
DaemonExtraArgs: defaultEnv("TS_TAILSCALED_EXTRA_ARGS", ""),
ExtraArgs: defaultEnv("TS_EXTRA_ARGS", ""),
InKubernetes: os.Getenv("KUBERNETES_SERVICE_HOST") != "",
UserspaceMode: defaultBool("TS_USERSPACE", true),
StateDir: defaultEnv("TS_STATE_DIR", ""),
AcceptDNS: defaultEnvBoolPointer("TS_ACCEPT_DNS"),
KubeSecret: defaultEnv("TS_KUBE_SECRET", "tailscale"),
SOCKSProxyAddr: defaultEnv("TS_SOCKS5_SERVER", ""),
HTTPProxyAddr: defaultEnv("TS_OUTBOUND_HTTP_PROXY_LISTEN", ""),
Socket: defaultEnv("TS_SOCKET", "/tmp/tailscaled.sock"),
AuthOnce: defaultBool("TS_AUTH_ONCE", false),
Root: defaultEnv("TS_TEST_ONLY_ROOT", "/"),
TailscaledConfigFilePath: defaultEnv("EXPERIMENTAL_TS_CONFIGFILE_PATH", ""),
AuthKey: defaultEnvs([]string{"TS_AUTHKEY", "TS_AUTH_KEY"}, ""),
Hostname: defaultEnv("TS_HOSTNAME", ""),
Routes: defaultEnvStringPointer("TS_ROUTES"),
ServeConfigPath: defaultEnv("TS_SERVE_CONFIG", ""),
ProxyTo: defaultEnv("TS_DEST_IP", ""),
TailnetTargetIP: defaultEnv("TS_TAILNET_TARGET_IP", ""),
TailnetTargetFQDN: defaultEnv("TS_TAILNET_TARGET_FQDN", ""),
DaemonExtraArgs: defaultEnv("TS_TAILSCALED_EXTRA_ARGS", ""),
ExtraArgs: defaultEnv("TS_EXTRA_ARGS", ""),
InKubernetes: os.Getenv("KUBERNETES_SERVICE_HOST") != "",
UserspaceMode: defaultBool("TS_USERSPACE", true),
StateDir: defaultEnv("TS_STATE_DIR", ""),
AcceptDNS: defaultEnvBoolPointer("TS_ACCEPT_DNS"),
KubeSecret: defaultEnv("TS_KUBE_SECRET", "tailscale"),
SOCKSProxyAddr: defaultEnv("TS_SOCKS5_SERVER", ""),
HTTPProxyAddr: defaultEnv("TS_OUTBOUND_HTTP_PROXY_LISTEN", ""),
Socket: defaultEnv("TS_SOCKET", "/tmp/tailscaled.sock"),
AuthOnce: defaultBool("TS_AUTH_ONCE", false),
Root: defaultEnv("TS_TEST_ONLY_ROOT", "/"),
TailscaledConfigFilePath: defaultEnv("EXPERIMENTAL_TS_CONFIGFILE_PATH", ""),
AllowProxyingClusterTrafficViaIngress: defaultBool("EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS", false),
PodIP: defaultEnv("POD_IP", ""),
}
if err := cfg.validate(); err != nil {
log.Fatalf("invalid configuration: %v", err)
}
@@ -330,7 +341,7 @@ authLoop:
}
var (
wantProxy = cfg.ProxyTo != "" || cfg.TailnetTargetIP != "" || cfg.TailnetTargetFQDN != ""
wantProxy = cfg.ProxyTo != "" || cfg.TailnetTargetIP != "" || cfg.TailnetTargetFQDN != "" || cfg.AllowProxyingClusterTrafficViaIngress
wantDeviceInfo = cfg.InKubernetes && cfg.KubeSecret != "" && cfg.KubernetesCanPatch
startupTasksDone = false
currentIPs deephash.Sum // tailscale IPs assigned to device
@@ -365,6 +376,7 @@ authLoop:
}
}()
var wg sync.WaitGroup
runLoop:
for {
select {
@@ -451,6 +463,18 @@ runLoop:
log.Fatalf("installing egress proxy rules: %v", err)
}
}
// If this is a L7 cluster ingress proxy (set up
// by Kubernetes operator) and proxying of
// cluster traffic to the ingress target is
// enabled, set up proxy rule each time the
// tailnet IPs of this node change (including
// the first time they become available).
if cfg.AllowProxyingClusterTrafficViaIngress && cfg.ServeConfigPath != "" && ipsHaveChanged && len(addrs) > 0 {
log.Printf("installing rules to forward traffic for %s to node's tailnet IP", cfg.PodIP)
if err := installTSForwardingRuleForDestination(ctx, cfg.PodIP, addrs, nfr); err != nil {
log.Fatalf("installing rules to forward traffic to node's tailnet IP: %v", err)
}
}
currentIPs = newCurrentIPs
deviceInfo := []any{n.NetMap.SelfNode.StableID(), n.NetMap.SelfNode.Name()}
@@ -837,6 +861,35 @@ func installEgressForwardingRule(ctx context.Context, dstStr string, tsIPs []net
return nil
}
// installTSForwardingRuleForDestination accepts a destination address and a
// list of node's tailnet addresses, sets up rules to forward traffic for
// destination to the tailnet IP matching the destination IP family.
// Destination can be Pod IP of this node.
func installTSForwardingRuleForDestination(ctx context.Context, dstFilter string, tsIPs []netip.Prefix, nfr linuxfw.NetfilterRunner) error {
dst, err := netip.ParseAddr(dstFilter)
if err != nil {
return err
}
var local netip.Addr
for _, pfx := range tsIPs {
if !pfx.IsSingleIP() {
continue
}
if pfx.Addr().Is4() != dst.Is4() {
continue
}
local = pfx.Addr()
break
}
if !local.IsValid() {
return fmt.Errorf("no tailscale IP matching family of %s found in %v", dstFilter, tsIPs)
}
if err := nfr.AddDNATRule(dst, local); err != nil {
return fmt.Errorf("installing rule for forwarding traffic to tailnet IP: %w", err)
}
return nil
}
func installIngressForwardingRule(ctx context.Context, dstStr string, tsIPs []netip.Prefix, nfr linuxfw.NetfilterRunner) error {
dst, err := netip.ParseAddr(dstStr)
if err != nil {
@@ -897,6 +950,14 @@ type settings struct {
Root string
KubernetesCanPatch bool
TailscaledConfigFilePath string
// If set to true and, if this containerboot instance is a Kubernetes
// ingress proxy, set up rules to forward incoming cluster traffic to be
// forwarded to the ingress target in cluster.
AllowProxyingClusterTrafficViaIngress bool
// PodIP is the IP of the Pod if running in Kubernetes. This is used
// when setting up rules to proxy cluster traffic to cluster ingress
// target.
PodIP string
}
func (s *settings) validate() error {
@@ -920,6 +981,15 @@ func (s *settings) validate() error {
if s.TailscaledConfigFilePath != "" && (s.AcceptDNS != nil || s.AuthKey != "" || s.Routes != nil || s.ExtraArgs != "" || s.Hostname != "") {
return errors.New("EXPERIMENTAL_TS_CONFIGFILE_PATH cannot be set in combination with TS_HOSTNAME, TS_EXTRA_ARGS, TS_AUTHKEY, TS_ROUTES, TS_ACCEPT_DNS.")
}
if s.AllowProxyingClusterTrafficViaIngress && s.UserspaceMode {
return errors.New("EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS is not supported in userspace mode")
}
if s.AllowProxyingClusterTrafficViaIngress && s.ServeConfigPath == "" {
return errors.New("EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS is set but this is not a cluster ingress proxy")
}
if s.AllowProxyingClusterTrafficViaIngress && s.PodIP == "" {
return errors.New("EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS is set but POD_IP is not set")
}
return nil
}

View File

@@ -11,7 +11,6 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
W 💣 github.com/dblohm7/wingoes from tailscale.com/util/winutil
github.com/fxamacker/cbor/v2 from tailscale.com/tka
github.com/golang/groupcache/lru from tailscale.com/net/dnscache
github.com/golang/protobuf/proto from github.com/matttproud/golang_protobuf_extensions/pbutil
L github.com/google/nftables from tailscale.com/util/linuxfw
L 💣 github.com/google/nftables/alignedbuff from github.com/google/nftables/xt
L 💣 github.com/google/nftables/binaryutil from github.com/google/nftables+
@@ -23,9 +22,7 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
L github.com/josharian/native from github.com/mdlayher/netlink+
L 💣 github.com/jsimonetti/rtnetlink from tailscale.com/net/interfaces+
L github.com/jsimonetti/rtnetlink/internal/unix from github.com/jsimonetti/rtnetlink
github.com/klauspost/compress/flate from nhooyr.io/websocket
github.com/matttproud/golang_protobuf_extensions/pbutil from github.com/prometheus/common/expfmt
L 💣 github.com/mdlayher/netlink from github.com/jsimonetti/rtnetlink+
L 💣 github.com/mdlayher/netlink from github.com/google/nftables+
L 💣 github.com/mdlayher/netlink/nlenc from github.com/jsimonetti/rtnetlink+
L github.com/mdlayher/netlink/nltest from github.com/google/nftables
L 💣 github.com/mdlayher/socket from github.com/mdlayher/netlink
@@ -49,10 +46,11 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
L github.com/vishvananda/netns from github.com/tailscale/netlink+
github.com/x448/float16 from github.com/fxamacker/cbor/v2
💣 go4.org/mem from tailscale.com/client/tailscale+
go4.org/netipx from tailscale.com/wgengine/filter+
go4.org/netipx from tailscale.com/net/tsaddr+
W 💣 golang.zx2c4.com/wireguard/windows/tunnel/winipcfg from tailscale.com/net/interfaces+
google.golang.org/protobuf/encoding/prototext from github.com/golang/protobuf/proto+
google.golang.org/protobuf/encoding/protowire from github.com/golang/protobuf/proto+
google.golang.org/protobuf/encoding/protodelim from github.com/prometheus/common/expfmt
google.golang.org/protobuf/encoding/prototext from github.com/prometheus/common/expfmt+
google.golang.org/protobuf/encoding/protowire from google.golang.org/protobuf/encoding/protodelim+
google.golang.org/protobuf/internal/descfmt from google.golang.org/protobuf/internal/filedesc
google.golang.org/protobuf/internal/descopts from google.golang.org/protobuf/internal/filedesc+
google.golang.org/protobuf/internal/detrand from google.golang.org/protobuf/internal/descfmt+
@@ -71,16 +69,15 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
google.golang.org/protobuf/internal/set from google.golang.org/protobuf/encoding/prototext
💣 google.golang.org/protobuf/internal/strs from google.golang.org/protobuf/encoding/prototext+
google.golang.org/protobuf/internal/version from google.golang.org/protobuf/runtime/protoimpl
google.golang.org/protobuf/proto from github.com/golang/protobuf/proto+
google.golang.org/protobuf/reflect/protodesc from github.com/golang/protobuf/proto
💣 google.golang.org/protobuf/reflect/protoreflect from github.com/golang/protobuf/proto+
google.golang.org/protobuf/reflect/protoregistry from github.com/golang/protobuf/proto+
google.golang.org/protobuf/runtime/protoiface from github.com/golang/protobuf/proto+
google.golang.org/protobuf/runtime/protoimpl from github.com/golang/protobuf/proto+
google.golang.org/protobuf/types/descriptorpb from google.golang.org/protobuf/reflect/protodesc
google.golang.org/protobuf/proto from github.com/prometheus/client_golang/prometheus+
💣 google.golang.org/protobuf/reflect/protoreflect from github.com/prometheus/client_model/go+
google.golang.org/protobuf/reflect/protoregistry from google.golang.org/protobuf/encoding/prototext+
google.golang.org/protobuf/runtime/protoiface from google.golang.org/protobuf/internal/impl+
google.golang.org/protobuf/runtime/protoimpl from github.com/prometheus/client_model/go+
google.golang.org/protobuf/types/known/timestamppb from github.com/prometheus/client_golang/prometheus+
nhooyr.io/websocket from tailscale.com/cmd/derper+
nhooyr.io/websocket/internal/errd from nhooyr.io/websocket
nhooyr.io/websocket/internal/util from nhooyr.io/websocket
nhooyr.io/websocket/internal/xsync from nhooyr.io/websocket
tailscale.com from tailscale.com/version
tailscale.com/atomicfile from tailscale.com/cmd/derper+
@@ -89,7 +86,7 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
tailscale.com/derp from tailscale.com/cmd/derper+
tailscale.com/derp/derphttp from tailscale.com/cmd/derper
tailscale.com/disco from tailscale.com/derp
tailscale.com/envknob from tailscale.com/derp+
tailscale.com/envknob from tailscale.com/client/tailscale+
tailscale.com/health from tailscale.com/net/tlsdial
tailscale.com/hostinfo from tailscale.com/net/interfaces+
tailscale.com/ipn from tailscale.com/client/tailscale
@@ -97,10 +94,11 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
tailscale.com/metrics from tailscale.com/cmd/derper+
tailscale.com/net/dnscache from tailscale.com/derp/derphttp
tailscale.com/net/flowtrack from tailscale.com/net/packet+
💣 tailscale.com/net/interfaces from tailscale.com/net/netns+
💣 tailscale.com/net/interfaces from tailscale.com/net/netmon+
tailscale.com/net/ktimeout from tailscale.com/cmd/derper
tailscale.com/net/netaddr from tailscale.com/ipn+
tailscale.com/net/netknob from tailscale.com/net/netns
tailscale.com/net/netmon from tailscale.com/net/sockstats+
tailscale.com/net/netmon from tailscale.com/derp/derphttp+
tailscale.com/net/netns from tailscale.com/derp/derphttp
tailscale.com/net/netutil from tailscale.com/client/tailscale
tailscale.com/net/packet from tailscale.com/wgengine/filter
@@ -116,18 +114,19 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
💣 tailscale.com/safesocket from tailscale.com/client/tailscale
tailscale.com/syncs from tailscale.com/cmd/derper+
tailscale.com/tailcfg from tailscale.com/client/tailscale+
tailscale.com/tailfs from tailscale.com/client/tailscale+
tailscale.com/tka from tailscale.com/client/tailscale+
W tailscale.com/tsconst from tailscale.com/net/interfaces
tailscale.com/tstime from tailscale.com/derp+
tailscale.com/tstime/mono from tailscale.com/tstime/rate
tailscale.com/tstime/rate from tailscale.com/wgengine/filter+
tailscale.com/tstime/rate from tailscale.com/derp+
tailscale.com/tsweb from tailscale.com/cmd/derper
tailscale.com/tsweb/promvarz from tailscale.com/tsweb
tailscale.com/tsweb/varz from tailscale.com/tsweb+
tailscale.com/types/dnstype from tailscale.com/tailcfg
tailscale.com/types/empty from tailscale.com/ipn
tailscale.com/types/ipproto from tailscale.com/net/flowtrack+
tailscale.com/types/key from tailscale.com/cmd/derper+
tailscale.com/types/key from tailscale.com/client/tailscale+
tailscale.com/types/lazy from tailscale.com/version+
tailscale.com/types/logger from tailscale.com/cmd/derper+
tailscale.com/types/netmap from tailscale.com/ipn
@@ -136,33 +135,33 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
tailscale.com/types/preftype from tailscale.com/ipn
tailscale.com/types/ptr from tailscale.com/hostinfo+
tailscale.com/types/structs from tailscale.com/ipn+
tailscale.com/types/tkatype from tailscale.com/types/key+
tailscale.com/types/views from tailscale.com/ipn/ipnstate+
tailscale.com/util/clientmetric from tailscale.com/net/tshttpproxy+
tailscale.com/types/tkatype from tailscale.com/client/tailscale+
tailscale.com/types/views from tailscale.com/ipn+
tailscale.com/util/clientmetric from tailscale.com/net/netmon+
tailscale.com/util/cloudenv from tailscale.com/hostinfo+
W tailscale.com/util/cmpver from tailscale.com/net/tshttpproxy
tailscale.com/util/cmpx from tailscale.com/cmd/derper+
tailscale.com/util/ctxkey from tailscale.com/tsweb+
L 💣 tailscale.com/util/dirwalk from tailscale.com/metrics
tailscale.com/util/dnsname from tailscale.com/hostinfo+
tailscale.com/util/httpm from tailscale.com/client/tailscale
tailscale.com/util/lineread from tailscale.com/hostinfo+
L tailscale.com/util/linuxfw from tailscale.com/net/netns
tailscale.com/util/mak from tailscale.com/syncs+
tailscale.com/util/mak from tailscale.com/net/interfaces+
tailscale.com/util/multierr from tailscale.com/health+
tailscale.com/util/nocasemaps from tailscale.com/types/ipproto
tailscale.com/util/set from tailscale.com/health+
tailscale.com/util/set from tailscale.com/derp+
tailscale.com/util/singleflight from tailscale.com/net/dnscache
tailscale.com/util/slicesx from tailscale.com/cmd/derper+
tailscale.com/util/syspolicy from tailscale.com/ipn
tailscale.com/util/vizerror from tailscale.com/tsweb+
tailscale.com/util/vizerror from tailscale.com/tailcfg+
W 💣 tailscale.com/util/winutil from tailscale.com/hostinfo+
tailscale.com/version from tailscale.com/derp+
tailscale.com/version/distro from tailscale.com/hostinfo+
tailscale.com/version/distro from tailscale.com/envknob+
tailscale.com/wgengine/filter from tailscale.com/types/netmap
golang.org/x/crypto/acme from golang.org/x/crypto/acme/autocert
golang.org/x/crypto/acme/autocert from tailscale.com/cmd/derper
golang.org/x/crypto/argon2 from tailscale.com/tka
golang.org/x/crypto/blake2b from golang.org/x/crypto/nacl/box+
golang.org/x/crypto/blake2b from golang.org/x/crypto/argon2+
golang.org/x/crypto/blake2s from tailscale.com/tka
golang.org/x/crypto/chacha20 from golang.org/x/crypto/chacha20poly1305
golang.org/x/crypto/chacha20poly1305 from crypto/tls
@@ -182,10 +181,10 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
golang.org/x/net/proxy from tailscale.com/net/netns
D golang.org/x/net/route from net+
golang.org/x/sync/errgroup from github.com/mdlayher/socket+
golang.org/x/sys/cpu from golang.org/x/crypto/blake2b+
LD golang.org/x/sys/unix from github.com/jsimonetti/rtnetlink/internal/unix+
W golang.org/x/sys/windows from golang.org/x/sys/windows/registry+
W golang.org/x/sys/windows/registry from golang.zx2c4.com/wireguard/windows/tunnel/winipcfg+
golang.org/x/sys/cpu from github.com/josharian/native+
LD golang.org/x/sys/unix from github.com/google/nftables+
W golang.org/x/sys/windows from github.com/dblohm7/wingoes+
W golang.org/x/sys/windows/registry from github.com/dblohm7/wingoes+
W golang.org/x/sys/windows/svc from golang.org/x/sys/windows/svc/mgr+
W golang.org/x/sys/windows/svc/mgr from tailscale.com/util/winutil
golang.org/x/text/secure/bidirule from golang.org/x/net/idna
@@ -197,10 +196,10 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
bytes from bufio+
cmp from slices+
compress/flate from compress/gzip+
compress/gzip from internal/profile+
compress/gzip from google.golang.org/protobuf/internal/impl+
container/list from crypto/tls+
context from crypto/tls+
crypto from crypto/ecdsa+
crypto from crypto/ecdh+
crypto/aes from crypto/ecdsa+
crypto/cipher from crypto/aes+
crypto/des from crypto/tls+
@@ -225,14 +224,14 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
embed from crypto/internal/nistec+
encoding from encoding/json+
encoding/asn1 from crypto/x509+
encoding/base32 from tailscale.com/tka+
encoding/base32 from github.com/fxamacker/cbor/v2+
encoding/base64 from encoding/json+
encoding/binary from compress/gzip+
encoding/hex from crypto/x509+
encoding/json from expvar+
encoding/pem from crypto/tls+
errors from bufio+
expvar from tailscale.com/cmd/derper+
expvar from github.com/prometheus/client_golang/prometheus+
flag from tailscale.com/cmd/derper+
fmt from compress/flate+
go/token from google.golang.org/protobuf/internal/strs
@@ -246,12 +245,12 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
io/ioutil from github.com/mitchellh/go-ps+
log from expvar+
log/internal from log
maps from tailscale.com/types/views+
maps from tailscale.com/ipn+
math from compress/flate+
math/big from crypto/dsa+
math/bits from compress/flate+
math/rand from github.com/mdlayher/netlink+
mime from mime/multipart+
mime from github.com/prometheus/common/expfmt+
mime/multipart from net/http
mime/quotedprintable from mime/multipart
net from crypto/tls+
@@ -263,15 +262,15 @@ tailscale.com/cmd/derper dependencies: (generated by github.com/tailscale/depawa
net/textproto from golang.org/x/net/http/httpguts+
net/url from crypto/x509+
os from crypto/rand+
os/exec from golang.zx2c4.com/wireguard/windows/tunnel/winipcfg+
os/exec from github.com/coreos/go-iptables/iptables+
os/signal from tailscale.com/cmd/derper
W os/user from tailscale.com/util/winutil
path from golang.org/x/crypto/acme/autocert+
path from github.com/prometheus/client_golang/prometheus/internal+
path/filepath from crypto/x509+
reflect from crypto/x509+
regexp from internal/profile+
regexp from github.com/coreos/go-iptables/iptables+
regexp/syntax from regexp
runtime/debug from golang.org/x/crypto/acme+
runtime/debug from github.com/prometheus/client_golang/prometheus+
runtime/metrics from github.com/prometheus/client_golang/prometheus+
runtime/pprof from net/http/pprof
runtime/trace from net/http/pprof+

View File

@@ -5,6 +5,7 @@
package main // import "tailscale.com/cmd/derper"
import (
"cmp"
"context"
"crypto/tls"
"encoding/json"
@@ -25,16 +26,16 @@ import (
"syscall"
"time"
"go4.org/mem"
"golang.org/x/time/rate"
"tailscale.com/atomicfile"
"tailscale.com/derp"
"tailscale.com/derp/derphttp"
"tailscale.com/metrics"
"tailscale.com/net/ktimeout"
"tailscale.com/net/stunserver"
"tailscale.com/tsweb"
"tailscale.com/types/key"
"tailscale.com/util/cmpx"
"tailscale.com/types/logger"
)
var (
@@ -49,14 +50,21 @@ var (
runSTUN = flag.Bool("stun", true, "whether to run a STUN server. It will bind to the same IP (if any) as the --addr flag value.")
runDERP = flag.Bool("derp", true, "whether to run a DERP server. The only reason to set this false is if you're decommissioning a server but want to keep its bootstrap DNS functionality still running.")
meshPSKFile = flag.String("mesh-psk-file", defaultMeshPSKFile(), "if non-empty, path to file containing the mesh pre-shared key file. It should contain some hex string; whitespace is trimmed.")
meshWith = flag.String("mesh-with", "", "optional comma-separated list of hostnames to mesh with; the server's own hostname can be in the list")
bootstrapDNS = flag.String("bootstrap-dns-names", "", "optional comma-separated list of hostnames to make available at /bootstrap-dns")
unpublishedDNS = flag.String("unpublished-bootstrap-dns-names", "", "optional comma-separated list of hostnames to make available at /bootstrap-dns and not publish in the list")
verifyClients = flag.Bool("verify-clients", false, "verify clients to this DERP server through a local tailscaled instance.")
meshPSKFile = flag.String("mesh-psk-file", defaultMeshPSKFile(), "if non-empty, path to file containing the mesh pre-shared key file. It should contain some hex string; whitespace is trimmed.")
meshWith = flag.String("mesh-with", "", "optional comma-separated list of hostnames to mesh with; the server's own hostname can be in the list")
bootstrapDNS = flag.String("bootstrap-dns-names", "", "optional comma-separated list of hostnames to make available at /bootstrap-dns")
unpublishedDNS = flag.String("unpublished-bootstrap-dns-names", "", "optional comma-separated list of hostnames to make available at /bootstrap-dns and not publish in the list")
verifyClients = flag.Bool("verify-clients", false, "verify clients to this DERP server through a local tailscaled instance.")
verifyClientURL = flag.String("verify-client-url", "", "if non-empty, an admission controller URL for permitting client connections; see tailcfg.DERPAdmitClientRequest")
verifyFailOpen = flag.Bool("verify-client-url-fail-open", true, "whether we fail open if --verify-client-url is unreachable")
acceptConnLimit = flag.Float64("accept-connection-limit", math.Inf(+1), "rate limit for accepting new connection")
acceptConnBurst = flag.Int("accept-connection-burst", math.MaxInt, "burst limit for accepting new connection")
// tcpKeepAlive is intentionally long, to reduce battery cost. There is an L7 keepalive on a higher frequency schedule.
tcpKeepAlive = flag.Duration("tcp-keepalive-time", 10*time.Minute, "TCP keepalive time")
// tcpUserTimeout is intentionally short, so that hung connections are cleaned up promptly. DERPs should be nearby users.
tcpUserTimeout = flag.Duration("tcp-user-timeout", 15*time.Second, "TCP user timeout")
)
var (
@@ -147,6 +155,8 @@ func main() {
s := derp.NewServer(cfg.PrivateKey, log.Printf)
s.SetVerifyClient(*verifyClients)
s.SetVerifyClientURL(*verifyClientURL)
s.SetVerifyClientURLFailOpen(*verifyFailOpen)
if *meshPSKFile != "" {
b, err := os.ReadFile(*meshPSKFile)
@@ -216,7 +226,16 @@ func main() {
}))
debug.Handle("traffic", "Traffic check", http.HandlerFunc(s.ServeDebugTraffic))
quietLogger := log.New(logFilter{}, "", 0)
// Longer lived DERP connections send an application layer keepalive. Note
// if the keepalive is hit, the user timeout will take precedence over the
// keepalive counter, so the probe if unanswered will take effect promptly,
// this is less tolerant of high loss, but high loss is unexpected.
lc := net.ListenConfig{
Control: ktimeout.UserTimeout(*tcpUserTimeout),
KeepAlive: *tcpKeepAlive,
}
quietLogger := log.New(logger.HTTPServerLogFilter{Inner: log.Printf}, "", 0)
httpsrv := &http.Server{
Addr: *addr,
Handler: mux,
@@ -292,7 +311,12 @@ func main() {
// duration exceeds server's WriteTimeout".
WriteTimeout: 5 * time.Minute,
}
err := port80srv.ListenAndServe()
ln, err := lc.Listen(context.Background(), "tcp", port80srv.Addr)
if err != nil {
log.Fatal(err)
}
defer ln.Close()
err = port80srv.Serve(ln)
if err != nil {
if err != http.ErrServerClosed {
log.Fatal(err)
@@ -300,10 +324,15 @@ func main() {
}
}()
}
err = rateLimitedListenAndServeTLS(httpsrv)
err = rateLimitedListenAndServeTLS(httpsrv, &lc)
} else {
log.Printf("derper: serving on %s", *addr)
err = httpsrv.ListenAndServe()
var ln net.Listener
ln, err = lc.Listen(context.Background(), "tcp", httpsrv.Addr)
if err != nil {
log.Fatal(err)
}
err = httpsrv.Serve(ln)
}
if err != nil && err != http.ErrServerClosed {
log.Fatalf("derper: %v", err)
@@ -368,8 +397,8 @@ func defaultMeshPSKFile() string {
return ""
}
func rateLimitedListenAndServeTLS(srv *http.Server) error {
ln, err := net.Listen("tcp", cmpx.Or(srv.Addr, ":https"))
func rateLimitedListenAndServeTLS(srv *http.Server, lc *net.ListenConfig) error {
ln, err := lc.Listen(context.Background(), "tcp", cmp.Or(srv.Addr, ":https"))
if err != nil {
return err
}
@@ -423,22 +452,3 @@ func (l *rateLimitedListener) Accept() (net.Conn, error) {
l.numAccepts.Add(1)
return cn, nil
}
// logFilter is used to filter out useless error logs that are logged to
// the net/http.Server.ErrorLog logger.
type logFilter struct{}
func (logFilter) Write(p []byte) (int, error) {
b := mem.B(p)
if mem.HasSuffix(b, mem.S(": EOF\n")) ||
mem.HasSuffix(b, mem.S(": i/o timeout\n")) ||
mem.HasSuffix(b, mem.S(": read: connection reset by peer\n")) ||
mem.HasSuffix(b, mem.S(": remote error: tls: bad certificate\n")) ||
mem.HasSuffix(b, mem.S(": tls: first record does not look like a TLS handshake\n")) {
// Skip this log message, but say that we processed it
return len(p), nil
}
log.Printf("%s", p)
return len(p), nil
}

View File

@@ -19,18 +19,31 @@ import (
)
var (
derpMapURL = flag.String("derp-map", "https://login.tailscale.com/derpmap/default", "URL to DERP map (https:// or file://)")
listen = flag.String("listen", ":8030", "HTTP listen address")
probeOnce = flag.Bool("once", false, "probe once and print results, then exit; ignores the listen flag")
spread = flag.Bool("spread", true, "whether to spread probing over time")
interval = flag.Duration("interval", 15*time.Second, "probe interval")
derpMapURL = flag.String("derp-map", "https://login.tailscale.com/derpmap/default", "URL to DERP map (https:// or file://)")
listen = flag.String("listen", ":8030", "HTTP listen address")
probeOnce = flag.Bool("once", false, "probe once and print results, then exit; ignores the listen flag")
spread = flag.Bool("spread", true, "whether to spread probing over time")
interval = flag.Duration("interval", 15*time.Second, "probe interval")
meshInterval = flag.Duration("mesh-interval", 15*time.Second, "mesh probe interval")
stunInterval = flag.Duration("stun-interval", 15*time.Second, "STUN probe interval")
tlsInterval = flag.Duration("tls-interval", 15*time.Second, "TLS probe interval")
bwInterval = flag.Duration("bw-interval", 0, "bandwidth probe interval (0 = no bandwidth probing)")
bwSize = flag.Int64("bw-probe-size-bytes", 1_000_000, "bandwidth probe size")
)
func main() {
flag.Parse()
p := prober.New().WithSpread(*spread).WithOnce(*probeOnce).WithMetricNamespace("derpprobe")
dp, err := prober.DERP(p, *derpMapURL, *interval, *interval, *interval)
opts := []prober.DERPOpt{
prober.WithMeshProbing(*meshInterval),
prober.WithSTUNProbing(*stunInterval),
prober.WithTLSProbing(*tlsInterval),
}
if *bwInterval > 0 {
opts = append(opts, prober.WithBandwidthProbing(*bwInterval, *bwSize))
}
dp, err := prober.DERP(p, *derpMapURL, opts...)
if err != nil {
log.Fatal(err)
}
@@ -53,6 +66,7 @@ func main() {
mux := http.NewServeMux()
tsweb.Debugger(mux)
mux.HandleFunc("/", http.HandlerFunc(serveFunc(p)))
log.Printf("Listening on %s", *listen)
log.Fatal(http.ListenAndServe(*listen, mux))
}

3
cmd/dist/dist.go vendored
View File

@@ -33,6 +33,9 @@ func getTargets() ([]dist.Target, error) {
// Since only we can provide packages to Synology for
// distribution, we default to building the "sideload" variant of
// packages that we distribute on pkgs.tailscale.com.
//
// To build for package center, run
// ./tool/go run ./cmd/dist build --synology-package-center synology
ret = append(ret, synology.Targets(synologyPackageCenter, nil)...)
return ret, nil
}

View File

@@ -7,6 +7,7 @@
package main
import (
"cmp"
"context"
"flag"
"fmt"
@@ -16,7 +17,6 @@ import (
"golang.org/x/oauth2/clientcredentials"
"tailscale.com/client/tailscale"
"tailscale.com/util/cmpx"
)
func main() {
@@ -40,7 +40,7 @@ func main() {
log.Fatal("at least one tag must be specified")
}
baseURL := cmpx.Or(os.Getenv("TS_BASE_URL"), "https://api.tailscale.com")
baseURL := cmp.Or(os.Getenv("TS_BASE_URL"), "https://api.tailscale.com")
credentials := clientcredentials.Config{
ClientID: clientID,

View File

@@ -158,11 +158,13 @@ func main() {
if !ok && (!oiok || !osok) {
log.Fatal("set envvar TS_API_KEY to your Tailscale API key or TS_OAUTH_ID and TS_OAUTH_SECRET to your Tailscale OAuth ID and Secret")
}
if ok && (oiok || osok) {
if apiKey != "" && (oauthId != "" || oauthSecret != "") {
log.Fatal("set either the envvar TS_API_KEY or TS_OAUTH_ID and TS_OAUTH_SECRET")
}
var client *http.Client
if oiok {
if oiok && (oauthId != "" || oauthSecret != "") {
// Both should ideally be set, but if either are non-empty it means the user had an intent
// to set _something_, so they should receive the oauth error flow.
oauthConfig := &clientcredentials.Config{
ClientID: oauthId,
ClientSecret: oauthSecret,

View File

@@ -31,10 +31,12 @@ var (
//go:embed hello.tmpl.html
var embeddedTemplate string
var localClient tailscale.LocalClient
func main() {
flag.Parse()
if *testIP != "" {
res, err := tailscale.WhoIs(context.Background(), *testIP)
res, err := localClient.WhoIs(context.Background(), *testIP)
if err != nil {
log.Fatal(err)
}
@@ -76,7 +78,7 @@ func main() {
GetCertificate: func(hi *tls.ClientHelloInfo) (*tls.Certificate, error) {
switch hi.ServerName {
case "hello.ts.net":
return tailscale.GetCertificate(hi)
return localClient.GetCertificate(hi)
case "hello.ipn.dev":
c, err := tls.LoadX509KeyPair(
"/etc/hello/hello.ipn.dev.crt",
@@ -170,7 +172,7 @@ func root(w http.ResponseWriter, r *http.Request) {
return
}
who, err := tailscale.WhoIs(r.Context(), r.RemoteAddr)
who, err := localClient.WhoIs(r.Context(), r.RemoteAddr)
var data tmplData
if err != nil {
if devMode() {

View File

@@ -430,6 +430,8 @@
<footer class="footer text-gray-600 text-center mb-12">
<p>Read about <a href="https://tailscale.com/kb/1017/install#advanced-features" class="text-blue-600 hover:text-blue-800"
target="_blank">what you can do next &rarr;</a></p>
<p>Read about <a href="https://tailscale.com/kb/1073/hello" class="text-blue-600 hover:text-blue-800"
target="_blank">the Hello service &rarr;</a></p>
</footer>
</main>
</body>

View File

@@ -164,6 +164,17 @@ func (a *ConnectorReconciler) maybeProvisionConnector(ctx context.Context, logge
hostname = string(cn.Spec.Hostname)
}
crl := childResourceLabels(cn.Name, a.tsnamespace, "connector")
proxyClass := cn.Spec.ProxyClass
if proxyClass != "" {
if ready, err := proxyClassIsReady(ctx, proxyClass, a.Client); err != nil {
return fmt.Errorf("error verifying ProxyClass for Connector: %w", err)
} else if !ready {
logger.Infof("ProxyClass %s specified for the Connector, but is not (yet) Ready, waiting..", proxyClass)
return nil
}
}
sts := &tailscaleSTSConfig{
ParentResourceName: cn.Name,
ParentResourceUID: string(cn.UID),
@@ -173,6 +184,7 @@ func (a *ConnectorReconciler) maybeProvisionConnector(ctx context.Context, logge
Connector: &connector{
isExitNode: cn.Spec.ExitNode,
},
ProxyClass: proxyClass,
}
if cn.Spec.SubnetRouter != nil && len(cn.Spec.SubnetRouter.AdvertiseRoutes) > 0 {

View File

@@ -67,45 +67,40 @@ func TestConnector(t *testing.T) {
fullName, shortName := findGenName(t, fc, "", "test", "connector")
opts := configOpts{
stsName: shortName,
secretName: fullName,
parentType: "connector",
hostname: "test-connector",
shouldUseDeclarativeConfig: true,
isExitNode: true,
subnetRoutes: "10.40.0.0/14",
confFileHash: "9321660203effb80983eaecc7b5ac5a8c53934926f46e895b9fe295dcfc5a904",
stsName: shortName,
secretName: fullName,
parentType: "connector",
hostname: "test-connector",
isExitNode: true,
subnetRoutes: "10.40.0.0/14",
}
expectEqual(t, fc, expectedSecret(t, opts))
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Add another route to be advertised.
mustUpdate[tsapi.Connector](t, fc, "", "test", func(conn *tsapi.Connector) {
conn.Spec.SubnetRouter.AdvertiseRoutes = []tsapi.Route{"10.40.0.0/14", "10.44.0.0/20"}
})
opts.subnetRoutes = "10.40.0.0/14,10.44.0.0/20"
opts.confFileHash = "fb6c4daf67425f983985750cd8d6f2beae77e614fcb34176604571f5623d6862"
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Remove a route.
mustUpdate[tsapi.Connector](t, fc, "", "test", func(conn *tsapi.Connector) {
conn.Spec.SubnetRouter.AdvertiseRoutes = []tsapi.Route{"10.44.0.0/20"}
})
opts.subnetRoutes = "10.44.0.0/20"
opts.confFileHash = "bacba177bcfe3849065cf6fee53d658a9bb4144197ac5b861727d69ea99742bb"
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Remove the subnet router.
mustUpdate[tsapi.Connector](t, fc, "", "test", func(conn *tsapi.Connector) {
conn.Spec.SubnetRouter = nil
})
opts.subnetRoutes = ""
opts.confFileHash = "7c421a99128eb80e79a285a82702f19f8f720615542a15bd794858a6275d8079"
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Re-add the subnet router.
mustUpdate[tsapi.Connector](t, fc, "", "test", func(conn *tsapi.Connector) {
@@ -114,9 +109,8 @@ func TestConnector(t *testing.T) {
}
})
opts.subnetRoutes = "10.44.0.0/20"
opts.confFileHash = "bacba177bcfe3849065cf6fee53d658a9bb4144197ac5b861727d69ea99742bb"
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Delete the Connector.
if err = fc.Delete(context.Background(), cn); err != nil {
@@ -152,25 +146,22 @@ func TestConnector(t *testing.T) {
fullName, shortName = findGenName(t, fc, "", "test", "connector")
opts = configOpts{
stsName: shortName,
secretName: fullName,
parentType: "connector",
shouldUseDeclarativeConfig: true,
subnetRoutes: "10.40.0.0/14",
hostname: "test-connector",
confFileHash: "57d922331890c9b1c8c6ae664394cb254334c551d9cd9db14537b5d9da9fb17e",
stsName: shortName,
secretName: fullName,
parentType: "connector",
subnetRoutes: "10.40.0.0/14",
hostname: "test-connector",
}
expectEqual(t, fc, expectedSecret(t, opts))
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Add an exit node.
mustUpdate[tsapi.Connector](t, fc, "", "test", func(conn *tsapi.Connector) {
conn.Spec.ExitNode = true
})
opts.isExitNode = true
opts.confFileHash = "1499b591fd97a50f0330db6ec09979792c49890cf31f5da5bb6a3f50dba1e77a"
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Delete the Connector.
if err = fc.Delete(context.Background(), cn); err != nil {
@@ -183,3 +174,103 @@ func TestConnector(t *testing.T) {
expectMissing[appsv1.StatefulSet](t, fc, "operator-ns", shortName)
expectMissing[corev1.Secret](t, fc, "operator-ns", fullName)
}
func TestConnectorWithProxyClass(t *testing.T) {
// Setup
pc := &tsapi.ProxyClass{
ObjectMeta: metav1.ObjectMeta{Name: "custom-metadata"},
Spec: tsapi.ProxyClassSpec{StatefulSet: &tsapi.StatefulSet{
Labels: map[string]string{"foo": "bar"},
Annotations: map[string]string{"bar.io/foo": "some-val"},
Pod: &tsapi.Pod{Annotations: map[string]string{"foo.io/bar": "some-val"}}}},
}
cn := &tsapi.Connector{
ObjectMeta: metav1.ObjectMeta{
Name: "test",
UID: types.UID("1234-UID"),
},
TypeMeta: metav1.TypeMeta{
Kind: tsapi.ConnectorKind,
APIVersion: "tailscale.io/v1alpha1",
},
Spec: tsapi.ConnectorSpec{
SubnetRouter: &tsapi.SubnetRouter{
AdvertiseRoutes: []tsapi.Route{"10.40.0.0/14"},
},
ExitNode: true,
},
}
fc := fake.NewClientBuilder().
WithScheme(tsapi.GlobalScheme).
WithObjects(pc, cn).
WithStatusSubresource(pc, cn).
Build()
ft := &fakeTSClient{}
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
cl := tstest.NewClock(tstest.ClockOpts{})
cr := &ConnectorReconciler{
Client: fc,
clock: cl,
ssr: &tailscaleSTSReconciler{
Client: fc,
tsClient: ft,
defaultTags: []string{"tag:k8s"},
operatorNamespace: "operator-ns",
proxyImage: "tailscale/tailscale",
},
logger: zl.Sugar(),
}
// 1. Connector is created with no ProxyClass specified, create
// resources with the default configuration.
expectReconciled(t, cr, "", "test")
fullName, shortName := findGenName(t, fc, "", "test", "connector")
opts := configOpts{
stsName: shortName,
secretName: fullName,
parentType: "connector",
hostname: "test-connector",
isExitNode: true,
subnetRoutes: "10.40.0.0/14",
}
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 2. Update Connector to specify a ProxyClass. ProxyClass is not yet
// ready, so its configuration is NOT applied to the Connector
// resources.
mustUpdate(t, fc, "", "test", func(conn *tsapi.Connector) {
conn.Spec.ProxyClass = "custom-metadata"
})
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 3. ProxyClass is set to Ready by proxy-class reconciler. Connector
// get reconciled and configuration from the ProxyClass is applied to
// its resources.
mustUpdateStatus(t, fc, "", "custom-metadata", func(pc *tsapi.ProxyClass) {
pc.Status = tsapi.ProxyClassStatus{
Conditions: []tsapi.ConnectorCondition{{
Status: metav1.ConditionTrue,
Type: tsapi.ProxyClassready,
ObservedGeneration: pc.Generation,
}}}
})
opts.proxyClass = pc.Name
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 4. Connector.spec.proxyClass field is unset, Connector gets
// reconciled and configuration from the ProxyClass is removed from the
// cluster resources for the Connector.
mustUpdate(t, fc, "", "test", func(conn *tsapi.Connector) {
conn.Spec.ProxyClass = ""
})
opts.proxyClass = ""
expectReconciled(t, cr, "", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
}

View File

@@ -49,6 +49,8 @@ spec:
image: {{ .Values.operatorConfig.image.repo }}{{- if .Values.operatorConfig.image.digest -}}{{ printf "@%s" .Values.operatorConfig.image.digest}}{{- else -}}{{ printf "%s" $operatorTag }}{{- end }}
imagePullPolicy: {{ .Values.operatorConfig.image.pullPolicy }}
env:
- name: OPERATOR_INITIAL_TAGS
value: {{ join "," .Values.operatorConfig.defaultTags }}
- name: OPERATOR_HOSTNAME
value: {{ .Values.operatorConfig.hostname }}
- name: OPERATOR_SECRET
@@ -59,8 +61,6 @@ spec:
valueFrom:
fieldRef:
fieldPath: metadata.namespace
- name: ENABLE_CONNECTOR
value: "{{ .Values.enableConnector }}"
- name: CLIENT_ID_FILE
value: /oauth/client_id
- name: CLIENT_SECRET_FILE

View File

@@ -0,0 +1,8 @@
apiVersion: networking.k8s.io/v1
kind: IngressClass
metadata:
name: tailscale # class name currently can not be changed
annotations: {} # we do not support default IngressClass annotation https://kubernetes.io/docs/concepts/services-networking/ingress/#default-ingress-class
spec:
controller: tailscale.com/ts-ingress # controller name currently can not be changed
# parameters: {} # currently no parameters are supported

View File

@@ -18,8 +18,11 @@ rules:
- apiGroups: ["networking.k8s.io"]
resources: ["ingresses", "ingresses/status"]
verbs: ["*"]
- apiGroups: ["networking.k8s.io"]
resources: ["ingressclasses"]
verbs: ["get", "list", "watch"]
- apiGroups: ["tailscale.com"]
resources: ["connectors", "connectors/status"]
resources: ["connectors", "connectors/status", "proxyclasses", "proxyclasses/status"]
verbs: ["get", "list", "watch", "update"]
---
apiVersion: rbac.authorization.k8s.io/v1

View File

@@ -8,13 +8,20 @@ oauth: {}
# clientId: ""
# clientSecret: ""
# enableConnector determines whether the operator should reconcile
# connector.tailscale.com custom resources. If set to true you have to install
# connector CRD in a separate step.
# You can do so by running 'kubectl apply -f ./cmd/k8s-operator/deploy/crds'.
enableConnector: "false"
# installCRDs determines whether tailscale.com CRDs should be installed as part
# of chart installation. We do not use Helm's CRD installation mechanism as that
# does not allow for upgrading CRDs.
# https://helm.sh/docs/chart_best_practices/custom_resource_definitions/
installCRDs: true
operatorConfig:
# ACL tag that operator will be tagged with. Operator must be made owner of
# these tags
# https://tailscale.com/kb/1236/kubernetes-operator/?q=operator#setting-up-the-kubernetes-operator
# Multiple tags are defined as array items and passed to the operator as a comma-separated string
defaultTags:
- "tag:k8s-operator"
image:
repo: tailscale/k8s-operator
# Digest will be prioritized over tag. If neither are set appVersion will be

View File

@@ -54,6 +54,9 @@ spec:
description: Hostname is the tailnet hostname that should be assigned to the Connector node. If unset, hostname defaults to <connector name>-connector. Hostname can contain lower case letters, numbers and dashes, it must not start or end with a dash and must be between 2 and 63 characters long.
type: string
pattern: ^[a-z0-9][a-z0-9-]{0,61}[a-z0-9]$
proxyClass:
description: ProxyClass is the name of the ProxyClass custom resource that contains configuration options that should be applied to the resources created for this Connector. If unset, the operator will create resources with the default configuration.
type: string
subnetRouter:
description: SubnetRouter defines subnet routes that the Connector node should expose to tailnet. If unset, none are exposed. https://tailscale.com/kb/1019/subnets/
type: object

View File

@@ -0,0 +1,494 @@
apiVersion: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
metadata:
annotations:
controller-gen.kubebuilder.io/version: v0.13.0
name: proxyclasses.tailscale.com
spec:
group: tailscale.com
names:
kind: ProxyClass
listKind: ProxyClassList
plural: proxyclasses
singular: proxyclass
scope: Cluster
versions:
- additionalPrinterColumns:
- description: Status of the ProxyClass.
jsonPath: .status.conditions[?(@.type == "ProxyClassReady")].reason
name: Status
type: string
name: v1alpha1
schema:
openAPIV3Schema:
type: object
required:
- spec
properties:
apiVersion:
description: 'APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
type: string
kind:
description: 'Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
type: string
metadata:
type: object
spec:
type: object
required:
- statefulSet
properties:
statefulSet:
description: Proxy's StatefulSet spec.
type: object
properties:
annotations:
description: Annotations that will be added to the StatefulSet created for the proxy. Any Annotations specified here will be merged with the default annotations applied to the StatefulSet by the Tailscale Kubernetes operator as well as any other annotations that might have been applied by other actors. Annotations must be valid Kubernetes annotations. https://kubernetes.io/docs/concepts/overview/working-with-objects/annotations/#syntax-and-character-set
type: object
additionalProperties:
type: string
labels:
description: Labels that will be added to the StatefulSet created for the proxy. Any labels specified here will be merged with the default labels applied to the StatefulSet by the Tailscale Kubernetes operator as well as any other labels that might have been applied by other actors. Label keys and values must be valid Kubernetes label keys and values. https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#syntax-and-character-set
type: object
additionalProperties:
type: string
pod:
description: Configuration for the proxy Pod.
type: object
properties:
annotations:
description: Annotations that will be added to the proxy Pod. Any annotations specified here will be merged with the default annotations applied to the Pod by the Tailscale Kubernetes operator. Annotations must be valid Kubernetes annotations. https://kubernetes.io/docs/concepts/overview/working-with-objects/annotations/#syntax-and-character-set
type: object
additionalProperties:
type: string
imagePullSecrets:
description: Proxy Pod's image pull Secrets. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#PodSpec
type: array
items:
description: LocalObjectReference contains enough information to let you locate the referenced object inside the same namespace.
type: object
properties:
name:
description: 'Name of the referent. More info: https://kubernetes.io/docs/concepts/overview/working-with-objects/names/#names TODO: Add other useful fields. apiVersion, kind, uid?'
type: string
x-kubernetes-map-type: atomic
labels:
description: Labels that will be added to the proxy Pod. Any labels specified here will be merged with the default labels applied to the Pod by the Tailscale Kubernetes operator. Label keys and values must be valid Kubernetes label keys and values. https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#syntax-and-character-set
type: object
additionalProperties:
type: string
nodeName:
description: Proxy Pod's node name. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling
type: string
nodeSelector:
description: Proxy Pod's node selector. By default Tailscale Kubernetes operator does not apply any node selector. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling
type: object
additionalProperties:
type: string
securityContext:
description: Proxy Pod's security context. By default Tailscale Kubernetes operator does not apply any Pod security context. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#security-context-2
type: object
properties:
fsGroup:
description: "A special supplemental group that applies to all containers in a pod. Some volume types allow the Kubelet to change the ownership of that volume to be owned by the pod: \n 1. The owning GID will be the FSGroup 2. The setgid bit is set (new files created in the volume will be owned by FSGroup) 3. The permission bits are OR'd with rw-rw---- \n If unset, the Kubelet will not modify the ownership and permissions of any volume. Note that this field cannot be set when spec.os.name is windows."
type: integer
format: int64
fsGroupChangePolicy:
description: 'fsGroupChangePolicy defines behavior of changing ownership and permission of the volume before being exposed inside Pod. This field will only apply to volume types which support fsGroup based ownership(and permissions). It will have no effect on ephemeral volume types such as: secret, configmaps and emptydir. Valid values are "OnRootMismatch" and "Always". If not specified, "Always" is used. Note that this field cannot be set when spec.os.name is windows.'
type: string
runAsGroup:
description: The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
type: integer
format: int64
runAsNonRoot:
description: Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: boolean
runAsUser:
description: The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
type: integer
format: int64
seLinuxOptions:
description: The SELinux context to be applied to all containers. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
type: object
properties:
level:
description: Level is SELinux level label that applies to the container.
type: string
role:
description: Role is a SELinux role label that applies to the container.
type: string
type:
description: Type is a SELinux type label that applies to the container.
type: string
user:
description: User is a SELinux user label that applies to the container.
type: string
seccompProfile:
description: The seccomp options to use by the containers in this pod. Note that this field cannot be set when spec.os.name is windows.
type: object
required:
- type
properties:
localhostProfile:
description: localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must be set if type is "Localhost". Must NOT be set for any other type.
type: string
type:
description: "type indicates which kind of seccomp profile will be applied. Valid options are: \n Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied."
type: string
supplementalGroups:
description: A list of groups applied to the first process run in each container, in addition to the container's primary GID, the fsGroup (if specified), and group memberships defined in the container image for the uid of the container process. If unspecified, no additional groups are added to any container. Note that group memberships defined in the container image for the uid of the container process are still effective, even if they are not included in this list. Note that this field cannot be set when spec.os.name is windows.
type: array
items:
type: integer
format: int64
sysctls:
description: Sysctls hold a list of namespaced sysctls used for the pod. Pods with unsupported sysctls (by the container runtime) might fail to launch. Note that this field cannot be set when spec.os.name is windows.
type: array
items:
description: Sysctl defines a kernel parameter to be set
type: object
required:
- name
- value
properties:
name:
description: Name of a property to set
type: string
value:
description: Value of a property to set
type: string
windowsOptions:
description: The Windows specific settings applied to all containers. If unspecified, the options within a container's SecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
type: object
properties:
gmsaCredentialSpec:
description: GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
type: string
gmsaCredentialSpecName:
description: GMSACredentialSpecName is the name of the GMSA credential spec to use.
type: string
hostProcess:
description: HostProcess determines if a container should be run as a 'Host Process' container. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
type: boolean
runAsUserName:
description: The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: string
tailscaleContainer:
description: Configuration for the proxy container running tailscale.
type: object
properties:
resources:
description: Container resource requirements. By default Tailscale Kubernetes operator does not apply any resource requirements. The amount of resources required wil depend on the amount of resources the operator needs to parse, usage patterns and cluster size. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#resources
type: object
properties:
claims:
description: "Claims lists the names of resources, defined in spec.resourceClaims, that are used by this container. \n This is an alpha field and requires enabling the DynamicResourceAllocation feature gate. \n This field is immutable. It can only be set for containers."
type: array
items:
description: ResourceClaim references one entry in PodSpec.ResourceClaims.
type: object
required:
- name
properties:
name:
description: Name must match the name of one entry in pod.spec.resourceClaims of the Pod where this field is used. It makes that resource available inside a container.
type: string
x-kubernetes-list-map-keys:
- name
x-kubernetes-list-type: map
limits:
description: 'Limits describes the maximum amount of compute resources allowed. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
additionalProperties:
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
anyOf:
- type: integer
- type: string
x-kubernetes-int-or-string: true
requests:
description: 'Requests describes the minimum amount of compute resources required. If Requests is omitted for a container, it defaults to Limits if that is explicitly specified, otherwise to an implementation-defined value. Requests cannot exceed Limits. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
additionalProperties:
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
anyOf:
- type: integer
- type: string
x-kubernetes-int-or-string: true
securityContext:
description: 'Container security context. Security context specified here will override the security context by the operator. By default the operator: - sets ''privileged: true'' for the init container - set NET_ADMIN capability for tailscale container for proxies that are created for Services or Connector. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#security-context'
type: object
properties:
allowPrivilegeEscalation:
description: 'AllowPrivilegeEscalation controls whether a process can gain more privileges than its parent process. This bool directly controls if the no_new_privs flag will be set on the container process. AllowPrivilegeEscalation is true always when the container is: 1) run as Privileged 2) has CAP_SYS_ADMIN Note that this field cannot be set when spec.os.name is windows.'
type: boolean
capabilities:
description: The capabilities to add/drop when running containers. Defaults to the default set of capabilities granted by the container runtime. Note that this field cannot be set when spec.os.name is windows.
type: object
properties:
add:
description: Added capabilities
type: array
items:
description: Capability represent POSIX capabilities type
type: string
drop:
description: Removed capabilities
type: array
items:
description: Capability represent POSIX capabilities type
type: string
privileged:
description: Run container in privileged mode. Processes in privileged containers are essentially equivalent to root on the host. Defaults to false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
procMount:
description: procMount denotes the type of proc mount to use for the containers. The default is DefaultProcMount which uses the container runtime defaults for readonly paths and masked paths. This requires the ProcMountType feature flag to be enabled. Note that this field cannot be set when spec.os.name is windows.
type: string
readOnlyRootFilesystem:
description: Whether this container has a read-only root filesystem. Default is false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
runAsGroup:
description: The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
type: integer
format: int64
runAsNonRoot:
description: Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: boolean
runAsUser:
description: The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
type: integer
format: int64
seLinuxOptions:
description: The SELinux context to be applied to the container. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
type: object
properties:
level:
description: Level is SELinux level label that applies to the container.
type: string
role:
description: Role is a SELinux role label that applies to the container.
type: string
type:
description: Type is a SELinux type label that applies to the container.
type: string
user:
description: User is a SELinux user label that applies to the container.
type: string
seccompProfile:
description: The seccomp options to use by this container. If seccomp options are provided at both the pod & container level, the container options override the pod options. Note that this field cannot be set when spec.os.name is windows.
type: object
required:
- type
properties:
localhostProfile:
description: localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must be set if type is "Localhost". Must NOT be set for any other type.
type: string
type:
description: "type indicates which kind of seccomp profile will be applied. Valid options are: \n Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied."
type: string
windowsOptions:
description: The Windows specific settings applied to all containers. If unspecified, the options from the PodSecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
type: object
properties:
gmsaCredentialSpec:
description: GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
type: string
gmsaCredentialSpecName:
description: GMSACredentialSpecName is the name of the GMSA credential spec to use.
type: string
hostProcess:
description: HostProcess determines if a container should be run as a 'Host Process' container. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
type: boolean
runAsUserName:
description: The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: string
tailscaleInitContainer:
description: Configuration for the proxy init container that enables forwarding.
type: object
properties:
resources:
description: Container resource requirements. By default Tailscale Kubernetes operator does not apply any resource requirements. The amount of resources required wil depend on the amount of resources the operator needs to parse, usage patterns and cluster size. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#resources
type: object
properties:
claims:
description: "Claims lists the names of resources, defined in spec.resourceClaims, that are used by this container. \n This is an alpha field and requires enabling the DynamicResourceAllocation feature gate. \n This field is immutable. It can only be set for containers."
type: array
items:
description: ResourceClaim references one entry in PodSpec.ResourceClaims.
type: object
required:
- name
properties:
name:
description: Name must match the name of one entry in pod.spec.resourceClaims of the Pod where this field is used. It makes that resource available inside a container.
type: string
x-kubernetes-list-map-keys:
- name
x-kubernetes-list-type: map
limits:
description: 'Limits describes the maximum amount of compute resources allowed. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
additionalProperties:
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
anyOf:
- type: integer
- type: string
x-kubernetes-int-or-string: true
requests:
description: 'Requests describes the minimum amount of compute resources required. If Requests is omitted for a container, it defaults to Limits if that is explicitly specified, otherwise to an implementation-defined value. Requests cannot exceed Limits. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
additionalProperties:
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
anyOf:
- type: integer
- type: string
x-kubernetes-int-or-string: true
securityContext:
description: 'Container security context. Security context specified here will override the security context by the operator. By default the operator: - sets ''privileged: true'' for the init container - set NET_ADMIN capability for tailscale container for proxies that are created for Services or Connector. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#security-context'
type: object
properties:
allowPrivilegeEscalation:
description: 'AllowPrivilegeEscalation controls whether a process can gain more privileges than its parent process. This bool directly controls if the no_new_privs flag will be set on the container process. AllowPrivilegeEscalation is true always when the container is: 1) run as Privileged 2) has CAP_SYS_ADMIN Note that this field cannot be set when spec.os.name is windows.'
type: boolean
capabilities:
description: The capabilities to add/drop when running containers. Defaults to the default set of capabilities granted by the container runtime. Note that this field cannot be set when spec.os.name is windows.
type: object
properties:
add:
description: Added capabilities
type: array
items:
description: Capability represent POSIX capabilities type
type: string
drop:
description: Removed capabilities
type: array
items:
description: Capability represent POSIX capabilities type
type: string
privileged:
description: Run container in privileged mode. Processes in privileged containers are essentially equivalent to root on the host. Defaults to false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
procMount:
description: procMount denotes the type of proc mount to use for the containers. The default is DefaultProcMount which uses the container runtime defaults for readonly paths and masked paths. This requires the ProcMountType feature flag to be enabled. Note that this field cannot be set when spec.os.name is windows.
type: string
readOnlyRootFilesystem:
description: Whether this container has a read-only root filesystem. Default is false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
runAsGroup:
description: The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
type: integer
format: int64
runAsNonRoot:
description: Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: boolean
runAsUser:
description: The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
type: integer
format: int64
seLinuxOptions:
description: The SELinux context to be applied to the container. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
type: object
properties:
level:
description: Level is SELinux level label that applies to the container.
type: string
role:
description: Role is a SELinux role label that applies to the container.
type: string
type:
description: Type is a SELinux type label that applies to the container.
type: string
user:
description: User is a SELinux user label that applies to the container.
type: string
seccompProfile:
description: The seccomp options to use by this container. If seccomp options are provided at both the pod & container level, the container options override the pod options. Note that this field cannot be set when spec.os.name is windows.
type: object
required:
- type
properties:
localhostProfile:
description: localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must be set if type is "Localhost". Must NOT be set for any other type.
type: string
type:
description: "type indicates which kind of seccomp profile will be applied. Valid options are: \n Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied."
type: string
windowsOptions:
description: The Windows specific settings applied to all containers. If unspecified, the options from the PodSecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
type: object
properties:
gmsaCredentialSpec:
description: GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
type: string
gmsaCredentialSpecName:
description: GMSACredentialSpecName is the name of the GMSA credential spec to use.
type: string
hostProcess:
description: HostProcess determines if a container should be run as a 'Host Process' container. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
type: boolean
runAsUserName:
description: The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: string
tolerations:
description: Proxy Pod's tolerations. By default Tailscale Kubernetes operator does not apply any tolerations. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling
type: array
items:
description: The pod this Toleration is attached to tolerates any taint that matches the triple <key,value,effect> using the matching operator <operator>.
type: object
properties:
effect:
description: Effect indicates the taint effect to match. Empty means match all taint effects. When specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.
type: string
key:
description: Key is the taint key that the toleration applies to. Empty means match all taint keys. If the key is empty, operator must be Exists; this combination means to match all values and all keys.
type: string
operator:
description: Operator represents a key's relationship to the value. Valid operators are Exists and Equal. Defaults to Equal. Exists is equivalent to wildcard for value, so that a pod can tolerate all taints of a particular category.
type: string
tolerationSeconds:
description: TolerationSeconds represents the period of time the toleration (which must be of effect NoExecute, otherwise this field is ignored) tolerates the taint. By default, it is not set, which means tolerate the taint forever (do not evict). Zero and negative values will be treated as 0 (evict immediately) by the system.
type: integer
format: int64
value:
description: Value is the taint value the toleration matches to. If the operator is Exists, the value should be empty, otherwise just a regular string.
type: string
status:
type: object
properties:
conditions:
description: List of status conditions to indicate the status of the ProxyClass. Known condition types are `ProxyClassReady`.
type: array
items:
description: ConnectorCondition contains condition information for a Connector.
type: object
required:
- status
- type
properties:
lastTransitionTime:
description: LastTransitionTime is the timestamp corresponding to the last status change of this condition.
type: string
format: date-time
message:
description: Message is a human readable description of the details of the last transition, complementing reason.
type: string
observedGeneration:
description: If set, this represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.condition[x].observedGeneration is 9, the condition is out of date with respect to the current state of the Connector.
type: integer
format: int64
reason:
description: Reason is a brief machine readable explanation for the condition's last transition.
type: string
status:
description: Status of the condition, one of ('True', 'False', 'Unknown').
type: string
type:
description: Type of the condition, known values are (`SubnetRouterReady`).
type: string
x-kubernetes-list-map-keys:
- type
x-kubernetes-list-type: map
served: true
storage: true
subresources:
status: {}

View File

@@ -0,0 +1,15 @@
apiVersion: tailscale.com/v1alpha1
kind: ProxyClass
metadata:
name: prod
spec:
statefulSet:
annotations:
platform-component: infra
pod:
labels:
team: eng
nodeSelector:
beta.kubernetes.io/os: "linux"
imagePullSecrets:
- name: "foo"

View File

@@ -27,6 +27,630 @@ metadata:
name: proxies
namespace: tailscale
---
apiVersion: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
metadata:
annotations:
controller-gen.kubebuilder.io/version: v0.13.0
name: connectors.tailscale.com
spec:
group: tailscale.com
names:
kind: Connector
listKind: ConnectorList
plural: connectors
shortNames:
- cn
singular: connector
scope: Cluster
versions:
- additionalPrinterColumns:
- description: CIDR ranges exposed to tailnet by a subnet router defined via this Connector instance.
jsonPath: .status.subnetRoutes
name: SubnetRoutes
type: string
- description: Whether this Connector instance defines an exit node.
jsonPath: .status.isExitNode
name: IsExitNode
type: string
- description: Status of the deployed Connector resources.
jsonPath: .status.conditions[?(@.type == "ConnectorReady")].reason
name: Status
type: string
name: v1alpha1
schema:
openAPIV3Schema:
properties:
apiVersion:
description: 'APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
type: string
kind:
description: 'Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
type: string
metadata:
type: object
spec:
description: ConnectorSpec describes the desired Tailscale component.
properties:
exitNode:
description: ExitNode defines whether the Connector node should act as a Tailscale exit node. Defaults to false. https://tailscale.com/kb/1103/exit-nodes
type: boolean
hostname:
description: Hostname is the tailnet hostname that should be assigned to the Connector node. If unset, hostname defaults to <connector name>-connector. Hostname can contain lower case letters, numbers and dashes, it must not start or end with a dash and must be between 2 and 63 characters long.
pattern: ^[a-z0-9][a-z0-9-]{0,61}[a-z0-9]$
type: string
proxyClass:
description: ProxyClass is the name of the ProxyClass custom resource that contains configuration options that should be applied to the resources created for this Connector. If unset, the operator will create resources with the default configuration.
type: string
subnetRouter:
description: SubnetRouter defines subnet routes that the Connector node should expose to tailnet. If unset, none are exposed. https://tailscale.com/kb/1019/subnets/
properties:
advertiseRoutes:
description: AdvertiseRoutes refer to CIDRs that the subnet router should make available. Route values must be strings that represent a valid IPv4 or IPv6 CIDR range. Values can be Tailscale 4via6 subnet routes. https://tailscale.com/kb/1201/4via6-subnets/
items:
format: cidr
type: string
minItems: 1
type: array
required:
- advertiseRoutes
type: object
tags:
description: Tags that the Tailscale node will be tagged with. Defaults to [tag:k8s]. To autoapprove the subnet routes or exit node defined by a Connector, you can configure Tailscale ACLs to give these tags the necessary permissions. See https://tailscale.com/kb/1018/acls/#auto-approvers-for-routes-and-exit-nodes. If you specify custom tags here, you must also make the operator an owner of these tags. See https://tailscale.com/kb/1236/kubernetes-operator/#setting-up-the-kubernetes-operator. Tags cannot be changed once a Connector node has been created. Tag values must be in form ^tag:[a-zA-Z][a-zA-Z0-9-]*$.
items:
pattern: ^tag:[a-zA-Z][a-zA-Z0-9-]*$
type: string
type: array
type: object
x-kubernetes-validations:
- message: A Connector needs to be either an exit node or a subnet router, or both.
rule: has(self.subnetRouter) || self.exitNode == true
status:
description: ConnectorStatus describes the status of the Connector. This is set and managed by the Tailscale operator.
properties:
conditions:
description: List of status conditions to indicate the status of the Connector. Known condition types are `ConnectorReady`.
items:
description: ConnectorCondition contains condition information for a Connector.
properties:
lastTransitionTime:
description: LastTransitionTime is the timestamp corresponding to the last status change of this condition.
format: date-time
type: string
message:
description: Message is a human readable description of the details of the last transition, complementing reason.
type: string
observedGeneration:
description: If set, this represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.condition[x].observedGeneration is 9, the condition is out of date with respect to the current state of the Connector.
format: int64
type: integer
reason:
description: Reason is a brief machine readable explanation for the condition's last transition.
type: string
status:
description: Status of the condition, one of ('True', 'False', 'Unknown').
type: string
type:
description: Type of the condition, known values are (`SubnetRouterReady`).
type: string
required:
- status
- type
type: object
type: array
x-kubernetes-list-map-keys:
- type
x-kubernetes-list-type: map
isExitNode:
description: IsExitNode is set to true if the Connector acts as an exit node.
type: boolean
subnetRoutes:
description: SubnetRoutes are the routes currently exposed to tailnet via this Connector instance.
type: string
type: object
required:
- spec
type: object
served: true
storage: true
subresources:
status: {}
---
apiVersion: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
metadata:
annotations:
controller-gen.kubebuilder.io/version: v0.13.0
name: proxyclasses.tailscale.com
spec:
group: tailscale.com
names:
kind: ProxyClass
listKind: ProxyClassList
plural: proxyclasses
singular: proxyclass
scope: Cluster
versions:
- additionalPrinterColumns:
- description: Status of the ProxyClass.
jsonPath: .status.conditions[?(@.type == "ProxyClassReady")].reason
name: Status
type: string
name: v1alpha1
schema:
openAPIV3Schema:
properties:
apiVersion:
description: 'APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
type: string
kind:
description: 'Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
type: string
metadata:
type: object
spec:
properties:
statefulSet:
description: Proxy's StatefulSet spec.
properties:
annotations:
additionalProperties:
type: string
description: Annotations that will be added to the StatefulSet created for the proxy. Any Annotations specified here will be merged with the default annotations applied to the StatefulSet by the Tailscale Kubernetes operator as well as any other annotations that might have been applied by other actors. Annotations must be valid Kubernetes annotations. https://kubernetes.io/docs/concepts/overview/working-with-objects/annotations/#syntax-and-character-set
type: object
labels:
additionalProperties:
type: string
description: Labels that will be added to the StatefulSet created for the proxy. Any labels specified here will be merged with the default labels applied to the StatefulSet by the Tailscale Kubernetes operator as well as any other labels that might have been applied by other actors. Label keys and values must be valid Kubernetes label keys and values. https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#syntax-and-character-set
type: object
pod:
description: Configuration for the proxy Pod.
properties:
annotations:
additionalProperties:
type: string
description: Annotations that will be added to the proxy Pod. Any annotations specified here will be merged with the default annotations applied to the Pod by the Tailscale Kubernetes operator. Annotations must be valid Kubernetes annotations. https://kubernetes.io/docs/concepts/overview/working-with-objects/annotations/#syntax-and-character-set
type: object
imagePullSecrets:
description: Proxy Pod's image pull Secrets. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#PodSpec
items:
description: LocalObjectReference contains enough information to let you locate the referenced object inside the same namespace.
properties:
name:
description: 'Name of the referent. More info: https://kubernetes.io/docs/concepts/overview/working-with-objects/names/#names TODO: Add other useful fields. apiVersion, kind, uid?'
type: string
type: object
x-kubernetes-map-type: atomic
type: array
labels:
additionalProperties:
type: string
description: Labels that will be added to the proxy Pod. Any labels specified here will be merged with the default labels applied to the Pod by the Tailscale Kubernetes operator. Label keys and values must be valid Kubernetes label keys and values. https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#syntax-and-character-set
type: object
nodeName:
description: Proxy Pod's node name. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling
type: string
nodeSelector:
additionalProperties:
type: string
description: Proxy Pod's node selector. By default Tailscale Kubernetes operator does not apply any node selector. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling
type: object
securityContext:
description: Proxy Pod's security context. By default Tailscale Kubernetes operator does not apply any Pod security context. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#security-context-2
properties:
fsGroup:
description: "A special supplemental group that applies to all containers in a pod. Some volume types allow the Kubelet to change the ownership of that volume to be owned by the pod: \n 1. The owning GID will be the FSGroup 2. The setgid bit is set (new files created in the volume will be owned by FSGroup) 3. The permission bits are OR'd with rw-rw---- \n If unset, the Kubelet will not modify the ownership and permissions of any volume. Note that this field cannot be set when spec.os.name is windows."
format: int64
type: integer
fsGroupChangePolicy:
description: 'fsGroupChangePolicy defines behavior of changing ownership and permission of the volume before being exposed inside Pod. This field will only apply to volume types which support fsGroup based ownership(and permissions). It will have no effect on ephemeral volume types such as: secret, configmaps and emptydir. Valid values are "OnRootMismatch" and "Always". If not specified, "Always" is used. Note that this field cannot be set when spec.os.name is windows.'
type: string
runAsGroup:
description: The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
format: int64
type: integer
runAsNonRoot:
description: Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: boolean
runAsUser:
description: The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
format: int64
type: integer
seLinuxOptions:
description: The SELinux context to be applied to all containers. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
properties:
level:
description: Level is SELinux level label that applies to the container.
type: string
role:
description: Role is a SELinux role label that applies to the container.
type: string
type:
description: Type is a SELinux type label that applies to the container.
type: string
user:
description: User is a SELinux user label that applies to the container.
type: string
type: object
seccompProfile:
description: The seccomp options to use by the containers in this pod. Note that this field cannot be set when spec.os.name is windows.
properties:
localhostProfile:
description: localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must be set if type is "Localhost". Must NOT be set for any other type.
type: string
type:
description: "type indicates which kind of seccomp profile will be applied. Valid options are: \n Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied."
type: string
required:
- type
type: object
supplementalGroups:
description: A list of groups applied to the first process run in each container, in addition to the container's primary GID, the fsGroup (if specified), and group memberships defined in the container image for the uid of the container process. If unspecified, no additional groups are added to any container. Note that group memberships defined in the container image for the uid of the container process are still effective, even if they are not included in this list. Note that this field cannot be set when spec.os.name is windows.
items:
format: int64
type: integer
type: array
sysctls:
description: Sysctls hold a list of namespaced sysctls used for the pod. Pods with unsupported sysctls (by the container runtime) might fail to launch. Note that this field cannot be set when spec.os.name is windows.
items:
description: Sysctl defines a kernel parameter to be set
properties:
name:
description: Name of a property to set
type: string
value:
description: Value of a property to set
type: string
required:
- name
- value
type: object
type: array
windowsOptions:
description: The Windows specific settings applied to all containers. If unspecified, the options within a container's SecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
properties:
gmsaCredentialSpec:
description: GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
type: string
gmsaCredentialSpecName:
description: GMSACredentialSpecName is the name of the GMSA credential spec to use.
type: string
hostProcess:
description: HostProcess determines if a container should be run as a 'Host Process' container. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
type: boolean
runAsUserName:
description: The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: string
type: object
type: object
tailscaleContainer:
description: Configuration for the proxy container running tailscale.
properties:
resources:
description: Container resource requirements. By default Tailscale Kubernetes operator does not apply any resource requirements. The amount of resources required wil depend on the amount of resources the operator needs to parse, usage patterns and cluster size. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#resources
properties:
claims:
description: "Claims lists the names of resources, defined in spec.resourceClaims, that are used by this container. \n This is an alpha field and requires enabling the DynamicResourceAllocation feature gate. \n This field is immutable. It can only be set for containers."
items:
description: ResourceClaim references one entry in PodSpec.ResourceClaims.
properties:
name:
description: Name must match the name of one entry in pod.spec.resourceClaims of the Pod where this field is used. It makes that resource available inside a container.
type: string
required:
- name
type: object
type: array
x-kubernetes-list-map-keys:
- name
x-kubernetes-list-type: map
limits:
additionalProperties:
anyOf:
- type: integer
- type: string
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
x-kubernetes-int-or-string: true
description: 'Limits describes the maximum amount of compute resources allowed. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
requests:
additionalProperties:
anyOf:
- type: integer
- type: string
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
x-kubernetes-int-or-string: true
description: 'Requests describes the minimum amount of compute resources required. If Requests is omitted for a container, it defaults to Limits if that is explicitly specified, otherwise to an implementation-defined value. Requests cannot exceed Limits. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
type: object
securityContext:
description: 'Container security context. Security context specified here will override the security context by the operator. By default the operator: - sets ''privileged: true'' for the init container - set NET_ADMIN capability for tailscale container for proxies that are created for Services or Connector. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#security-context'
properties:
allowPrivilegeEscalation:
description: 'AllowPrivilegeEscalation controls whether a process can gain more privileges than its parent process. This bool directly controls if the no_new_privs flag will be set on the container process. AllowPrivilegeEscalation is true always when the container is: 1) run as Privileged 2) has CAP_SYS_ADMIN Note that this field cannot be set when spec.os.name is windows.'
type: boolean
capabilities:
description: The capabilities to add/drop when running containers. Defaults to the default set of capabilities granted by the container runtime. Note that this field cannot be set when spec.os.name is windows.
properties:
add:
description: Added capabilities
items:
description: Capability represent POSIX capabilities type
type: string
type: array
drop:
description: Removed capabilities
items:
description: Capability represent POSIX capabilities type
type: string
type: array
type: object
privileged:
description: Run container in privileged mode. Processes in privileged containers are essentially equivalent to root on the host. Defaults to false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
procMount:
description: procMount denotes the type of proc mount to use for the containers. The default is DefaultProcMount which uses the container runtime defaults for readonly paths and masked paths. This requires the ProcMountType feature flag to be enabled. Note that this field cannot be set when spec.os.name is windows.
type: string
readOnlyRootFilesystem:
description: Whether this container has a read-only root filesystem. Default is false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
runAsGroup:
description: The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
format: int64
type: integer
runAsNonRoot:
description: Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: boolean
runAsUser:
description: The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
format: int64
type: integer
seLinuxOptions:
description: The SELinux context to be applied to the container. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
properties:
level:
description: Level is SELinux level label that applies to the container.
type: string
role:
description: Role is a SELinux role label that applies to the container.
type: string
type:
description: Type is a SELinux type label that applies to the container.
type: string
user:
description: User is a SELinux user label that applies to the container.
type: string
type: object
seccompProfile:
description: The seccomp options to use by this container. If seccomp options are provided at both the pod & container level, the container options override the pod options. Note that this field cannot be set when spec.os.name is windows.
properties:
localhostProfile:
description: localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must be set if type is "Localhost". Must NOT be set for any other type.
type: string
type:
description: "type indicates which kind of seccomp profile will be applied. Valid options are: \n Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied."
type: string
required:
- type
type: object
windowsOptions:
description: The Windows specific settings applied to all containers. If unspecified, the options from the PodSecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
properties:
gmsaCredentialSpec:
description: GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
type: string
gmsaCredentialSpecName:
description: GMSACredentialSpecName is the name of the GMSA credential spec to use.
type: string
hostProcess:
description: HostProcess determines if a container should be run as a 'Host Process' container. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
type: boolean
runAsUserName:
description: The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: string
type: object
type: object
type: object
tailscaleInitContainer:
description: Configuration for the proxy init container that enables forwarding.
properties:
resources:
description: Container resource requirements. By default Tailscale Kubernetes operator does not apply any resource requirements. The amount of resources required wil depend on the amount of resources the operator needs to parse, usage patterns and cluster size. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#resources
properties:
claims:
description: "Claims lists the names of resources, defined in spec.resourceClaims, that are used by this container. \n This is an alpha field and requires enabling the DynamicResourceAllocation feature gate. \n This field is immutable. It can only be set for containers."
items:
description: ResourceClaim references one entry in PodSpec.ResourceClaims.
properties:
name:
description: Name must match the name of one entry in pod.spec.resourceClaims of the Pod where this field is used. It makes that resource available inside a container.
type: string
required:
- name
type: object
type: array
x-kubernetes-list-map-keys:
- name
x-kubernetes-list-type: map
limits:
additionalProperties:
anyOf:
- type: integer
- type: string
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
x-kubernetes-int-or-string: true
description: 'Limits describes the maximum amount of compute resources allowed. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
requests:
additionalProperties:
anyOf:
- type: integer
- type: string
pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
x-kubernetes-int-or-string: true
description: 'Requests describes the minimum amount of compute resources required. If Requests is omitted for a container, it defaults to Limits if that is explicitly specified, otherwise to an implementation-defined value. Requests cannot exceed Limits. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
type: object
type: object
securityContext:
description: 'Container security context. Security context specified here will override the security context by the operator. By default the operator: - sets ''privileged: true'' for the init container - set NET_ADMIN capability for tailscale container for proxies that are created for Services or Connector. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#security-context'
properties:
allowPrivilegeEscalation:
description: 'AllowPrivilegeEscalation controls whether a process can gain more privileges than its parent process. This bool directly controls if the no_new_privs flag will be set on the container process. AllowPrivilegeEscalation is true always when the container is: 1) run as Privileged 2) has CAP_SYS_ADMIN Note that this field cannot be set when spec.os.name is windows.'
type: boolean
capabilities:
description: The capabilities to add/drop when running containers. Defaults to the default set of capabilities granted by the container runtime. Note that this field cannot be set when spec.os.name is windows.
properties:
add:
description: Added capabilities
items:
description: Capability represent POSIX capabilities type
type: string
type: array
drop:
description: Removed capabilities
items:
description: Capability represent POSIX capabilities type
type: string
type: array
type: object
privileged:
description: Run container in privileged mode. Processes in privileged containers are essentially equivalent to root on the host. Defaults to false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
procMount:
description: procMount denotes the type of proc mount to use for the containers. The default is DefaultProcMount which uses the container runtime defaults for readonly paths and masked paths. This requires the ProcMountType feature flag to be enabled. Note that this field cannot be set when spec.os.name is windows.
type: string
readOnlyRootFilesystem:
description: Whether this container has a read-only root filesystem. Default is false. Note that this field cannot be set when spec.os.name is windows.
type: boolean
runAsGroup:
description: The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
format: int64
type: integer
runAsNonRoot:
description: Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: boolean
runAsUser:
description: The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
format: int64
type: integer
seLinuxOptions:
description: The SELinux context to be applied to the container. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
properties:
level:
description: Level is SELinux level label that applies to the container.
type: string
role:
description: Role is a SELinux role label that applies to the container.
type: string
type:
description: Type is a SELinux type label that applies to the container.
type: string
user:
description: User is a SELinux user label that applies to the container.
type: string
type: object
seccompProfile:
description: The seccomp options to use by this container. If seccomp options are provided at both the pod & container level, the container options override the pod options. Note that this field cannot be set when spec.os.name is windows.
properties:
localhostProfile:
description: localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must be set if type is "Localhost". Must NOT be set for any other type.
type: string
type:
description: "type indicates which kind of seccomp profile will be applied. Valid options are: \n Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied."
type: string
required:
- type
type: object
windowsOptions:
description: The Windows specific settings applied to all containers. If unspecified, the options from the PodSecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
properties:
gmsaCredentialSpec:
description: GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
type: string
gmsaCredentialSpecName:
description: GMSACredentialSpecName is the name of the GMSA credential spec to use.
type: string
hostProcess:
description: HostProcess determines if a container should be run as a 'Host Process' container. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
type: boolean
runAsUserName:
description: The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
type: string
type: object
type: object
type: object
tolerations:
description: Proxy Pod's tolerations. By default Tailscale Kubernetes operator does not apply any tolerations. https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling
items:
description: The pod this Toleration is attached to tolerates any taint that matches the triple <key,value,effect> using the matching operator <operator>.
properties:
effect:
description: Effect indicates the taint effect to match. Empty means match all taint effects. When specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.
type: string
key:
description: Key is the taint key that the toleration applies to. Empty means match all taint keys. If the key is empty, operator must be Exists; this combination means to match all values and all keys.
type: string
operator:
description: Operator represents a key's relationship to the value. Valid operators are Exists and Equal. Defaults to Equal. Exists is equivalent to wildcard for value, so that a pod can tolerate all taints of a particular category.
type: string
tolerationSeconds:
description: TolerationSeconds represents the period of time the toleration (which must be of effect NoExecute, otherwise this field is ignored) tolerates the taint. By default, it is not set, which means tolerate the taint forever (do not evict). Zero and negative values will be treated as 0 (evict immediately) by the system.
format: int64
type: integer
value:
description: Value is the taint value the toleration matches to. If the operator is Exists, the value should be empty, otherwise just a regular string.
type: string
type: object
type: array
type: object
type: object
required:
- statefulSet
type: object
status:
properties:
conditions:
description: List of status conditions to indicate the status of the ProxyClass. Known condition types are `ProxyClassReady`.
items:
description: ConnectorCondition contains condition information for a Connector.
properties:
lastTransitionTime:
description: LastTransitionTime is the timestamp corresponding to the last status change of this condition.
format: date-time
type: string
message:
description: Message is a human readable description of the details of the last transition, complementing reason.
type: string
observedGeneration:
description: If set, this represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.condition[x].observedGeneration is 9, the condition is out of date with respect to the current state of the Connector.
format: int64
type: integer
reason:
description: Reason is a brief machine readable explanation for the condition's last transition.
type: string
status:
description: Status of the condition, one of ('True', 'False', 'Unknown').
type: string
type:
description: Type of the condition, known values are (`SubnetRouterReady`).
type: string
required:
- status
- type
type: object
type: array
x-kubernetes-list-map-keys:
- type
x-kubernetes-list-type: map
type: object
required:
- spec
type: object
served: true
storage: true
subresources:
status: {}
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
@@ -47,11 +671,21 @@ rules:
- ingresses/status
verbs:
- '*'
- apiGroups:
- networking.k8s.io
resources:
- ingressclasses
verbs:
- get
- list
- watch
- apiGroups:
- tailscale.com
resources:
- connectors
- connectors/status
- proxyclasses
- proxyclasses/status
verbs:
- get
- list
@@ -150,6 +784,8 @@ spec:
spec:
containers:
- env:
- name: OPERATOR_INITIAL_TAGS
value: tag:k8s-operator
- name: OPERATOR_HOSTNAME
value: tailscale-operator
- name: OPERATOR_SECRET
@@ -160,8 +796,6 @@ spec:
valueFrom:
fieldRef:
fieldPath: metadata.namespace
- name: ENABLE_CONNECTOR
value: "false"
- name: CLIENT_ID_FILE
value: /oauth/client_id
- name: CLIENT_SECRET_FILE
@@ -188,3 +822,11 @@ spec:
- name: oauth
secret:
secretName: operator-oauth
---
apiVersion: networking.k8s.io/v1
kind: IngressClass
metadata:
annotations: {}
name: tailscale
spec:
controller: tailscale.com/ts-ingress

View File

@@ -28,8 +28,10 @@ spec:
env:
- name: TS_USERSPACE
value: "false"
- name: TS_AUTH_ONCE
value: "true"
- name: POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
securityContext:
capabilities:
add:

View File

@@ -20,5 +20,3 @@ spec:
env:
- name: TS_USERSPACE
value: "true"
- name: TS_AUTH_ONCE
value: "true"

View File

@@ -18,15 +18,51 @@ import (
"gopkg.in/yaml.v3"
)
const (
operatorDeploymentFilesPath = "cmd/k8s-operator/deploy"
connectorCRDPath = operatorDeploymentFilesPath + "/crds/tailscale.com_connectors.yaml"
proxyClassCRDPath = operatorDeploymentFilesPath + "/crds/tailscale.com_proxyclasses.yaml"
helmTemplatesPath = operatorDeploymentFilesPath + "/chart/templates"
connectorCRDHelmTemplatePath = helmTemplatesPath + "/connector.yaml"
proxyClassCRDHelmTemplatePath = helmTemplatesPath + "/proxyclass.yaml"
helmConditionalStart = "{{ if .Values.installCRDs -}}\n"
helmConditionalEnd = "{{- end -}}"
)
func main() {
if len(os.Args) < 2 {
log.Fatalf("usage ./generate [staticmanifests|helmcrd]")
}
repoRoot := "../../"
cmd := exec.Command("./tool/helm", "template", "operator", "./cmd/k8s-operator/deploy/chart",
switch os.Args[1] {
case "helmcrd": // insert CRD to Helm templates behind a installCRDs=true conditional check
log.Print("Adding Connector CRD to Helm templates")
if err := generate("./"); err != nil {
log.Fatalf("error adding Connector CRD to Helm templates: %v", err)
}
return
case "staticmanifests": // generate static manifests from Helm templates (including the CRD)
default:
log.Fatalf("unknown option %s, known options are 'staticmanifests', 'helmcrd'", os.Args[1])
}
log.Printf("Inserting CRDs Helm templates")
if err := generate(repoRoot); err != nil {
log.Fatalf("error adding CRDs to Helm templates: %v", err)
}
defer func() {
if err := cleanup(repoRoot); err != nil {
log.Fatalf("error cleaning up generated resources")
}
}()
log.Print("Templating Helm chart contents")
helmTmplCmd := exec.Command("./tool/helm", "template", "operator", "./cmd/k8s-operator/deploy/chart",
"--namespace=tailscale")
cmd.Dir = repoRoot
helmTmplCmd.Dir = repoRoot
var out bytes.Buffer
cmd.Stdout = &out
cmd.Stderr = os.Stderr
if err := cmd.Run(); err != nil {
helmTmplCmd.Stdout = &out
helmTmplCmd.Stderr = os.Stderr
if err := helmTmplCmd.Run(); err != nil {
log.Fatalf("error templating helm manifests: %v", err)
}
@@ -54,7 +90,6 @@ func main() {
if err != nil {
log.Fatalf("failed read from input data: %v", err)
}
bytes, err := yaml.Marshal(document)
if err != nil {
log.Fatalf("failed to marshal YAML document: %v", err)
@@ -72,3 +107,49 @@ func main() {
log.Fatalf("error writing new file: %v", err)
}
}
// generate places tailscale.com CRDs (currently Connector and ProxyClass) into
// the Helm chart templates behind .Values.installCRDs=true condition (true by
// default).
func generate(baseDir string) error {
addCRDToHelm := func(crdPath, crdTemplatePath string) error {
chartBytes, err := os.ReadFile(filepath.Join(baseDir, crdPath))
if err != nil {
return fmt.Errorf("error reading CRD contents: %w", err)
}
// Place a new temporary Helm template file with the templated CRD
// contents into Helm templates.
file, err := os.Create(filepath.Join(baseDir, crdTemplatePath))
if err != nil {
return fmt.Errorf("error creating CRD template file: %w", err)
}
if _, err := file.Write([]byte(helmConditionalStart)); err != nil {
return fmt.Errorf("error writing helm if statement start: %w", err)
}
if _, err := file.Write(chartBytes); err != nil {
return fmt.Errorf("error writing chart bytes: %w", err)
}
if _, err := file.Write([]byte(helmConditionalEnd)); err != nil {
return fmt.Errorf("error writing helm if-statement end: %w", err)
}
return nil
}
if err := addCRDToHelm(connectorCRDPath, connectorCRDHelmTemplatePath); err != nil {
return fmt.Errorf("error adding Connector CRD to Helm templates: %w", err)
}
if err := addCRDToHelm(proxyClassCRDPath, proxyClassCRDHelmTemplatePath); err != nil {
return fmt.Errorf("error adding ProxyClass CRD to Helm templates: %w", err)
}
return nil
}
func cleanup(baseDir string) error {
log.Print("Cleaning up CRD from Helm templates")
if err := os.Remove(filepath.Join(baseDir, connectorCRDHelmTemplatePath)); err != nil && !os.IsNotExist(err) {
return fmt.Errorf("error cleaning up Connector CRD template: %w", err)
}
if err := os.Remove(filepath.Join(baseDir, proxyClassCRDHelmTemplatePath)); err != nil && !os.IsNotExist(err) {
return fmt.Errorf("error cleaning up ProxyClass CRD template: %w", err)
}
return nil
}

View File

@@ -0,0 +1,74 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
//go:build !plan9 && !windows
package main
import (
"bytes"
"os"
"os/exec"
"path/filepath"
"strings"
"testing"
)
func Test_generate(t *testing.T) {
base, err := os.Getwd()
base = filepath.Join(base, "../../../")
if err != nil {
t.Fatalf("error getting current working directory: %v", err)
}
defer cleanup(base)
if err := generate(base); err != nil {
t.Fatalf("CRD template generation: %v", err)
}
tempDir := t.TempDir()
helmCLIPath := filepath.Join(base, "tool/helm")
helmChartTemplatesPath := filepath.Join(base, "cmd/k8s-operator/deploy/chart")
helmPackageCmd := exec.Command(helmCLIPath, "package", helmChartTemplatesPath, "--destination", tempDir, "--version", "0.0.1")
helmPackageCmd.Stderr = os.Stderr
helmPackageCmd.Stdout = os.Stdout
if err := helmPackageCmd.Run(); err != nil {
t.Fatalf("error packaging Helm chart: %v", err)
}
helmPackagePath := filepath.Join(tempDir, "tailscale-operator-0.0.1.tgz")
helmLintCmd := exec.Command(helmCLIPath, "lint", helmPackagePath)
helmLintCmd.Stderr = os.Stderr
helmLintCmd.Stdout = os.Stdout
if err := helmLintCmd.Run(); err != nil {
t.Fatalf("Helm chart linter failed: %v", err)
}
// Test that default Helm install contains the Connector and ProxyClass CRDs.
installContentsWithCRD := bytes.NewBuffer([]byte{})
helmTemplateWithCRDCmd := exec.Command(helmCLIPath, "template", helmPackagePath)
helmTemplateWithCRDCmd.Stderr = os.Stderr
helmTemplateWithCRDCmd.Stdout = installContentsWithCRD
if err := helmTemplateWithCRDCmd.Run(); err != nil {
t.Fatalf("templating Helm chart with CRDs failed: %v", err)
}
if !strings.Contains(installContentsWithCRD.String(), "name: connectors.tailscale.com") {
t.Errorf("Connector CRD not found in default chart install")
}
if !strings.Contains(installContentsWithCRD.String(), "name: proxyclasses.tailscale.com") {
t.Errorf("ProxyClass CRD not found in default chart install")
}
// Test that CRDs can be excluded from Helm chart install
installContentsWithoutCRD := bytes.NewBuffer([]byte{})
helmTemplateWithoutCRDCmd := exec.Command(helmCLIPath, "template", helmPackagePath, "--set", "installCRDs=false")
helmTemplateWithoutCRDCmd.Stderr = os.Stderr
helmTemplateWithoutCRDCmd.Stdout = installContentsWithoutCRD
if err := helmTemplateWithoutCRDCmd.Run(); err != nil {
t.Fatalf("templating Helm chart without CRDs failed: %v", err)
}
if strings.Contains(installContentsWithoutCRD.String(), "name: connectors.tailscale.com") {
t.Errorf("Connector CRD found in chart install that should not contain a CRD")
}
if strings.Contains(installContentsWithoutCRD.String(), "name: connectors.tailscale.com") {
t.Errorf("ProxyClass CRD found in chart install that should not contain a CRD")
}
}

View File

@@ -12,10 +12,12 @@ import (
"strings"
"sync"
"github.com/pkg/errors"
"go.uber.org/zap"
corev1 "k8s.io/api/core/v1"
networkingv1 "k8s.io/api/networking/v1"
apierrors "k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"k8s.io/client-go/tools/record"
"sigs.k8s.io/controller-runtime/pkg/client"
@@ -26,6 +28,12 @@ import (
"tailscale.com/util/set"
)
const (
tailscaleIngressClassName = "tailscale" // ingressClass.metadata.name for tailscale IngressClass resource
tailscaleIngressControllerName = "tailscale.com/ts-ingress" // ingressClass.spec.controllerName for tailscale IngressClass resource
ingressClassDefaultAnnotation = "ingressclass.kubernetes.io/is-default-class" // we do not support this https://kubernetes.io/docs/concepts/services-networking/ingress/#default-ingress-class
)
type IngressReconciler struct {
client.Client
@@ -109,6 +117,10 @@ func (a *IngressReconciler) maybeCleanup(ctx context.Context, logger *zap.Sugare
// This function adds a finalizer to ing, ensuring that we can handle orderly
// deprovisioning later.
func (a *IngressReconciler) maybeProvision(ctx context.Context, logger *zap.SugaredLogger, ing *networkingv1.Ingress) error {
if err := a.validateIngressClass(ctx); err != nil {
logger.Warnf("error validating tailscale IngressClass: %v. In future this might be a terminal error.", err)
}
if !slices.Contains(ing.Finalizers, FinalizerName) {
// This log line is printed exactly once during initial provisioning,
// because once the finalizer is in place this block gets skipped. So,
@@ -120,6 +132,17 @@ func (a *IngressReconciler) maybeProvision(ctx context.Context, logger *zap.Suga
return fmt.Errorf("failed to add finalizer: %w", err)
}
}
proxyClass := proxyClassForObject(ing)
if proxyClass != "" {
if ready, err := proxyClassIsReady(ctx, proxyClass, a.Client); err != nil {
return fmt.Errorf("error verifying ProxyClass for Ingress: %w", err)
} else if !ready {
logger.Infof("ProxyClass %s specified for the Ingress, but is not (yet) Ready, waiting..", proxyClass)
return nil
}
}
a.mu.Lock()
a.managedIngresses.Add(ing.UID)
gaugeIngressResources.Set(int64(a.managedIngresses.Len()))
@@ -205,10 +228,25 @@ func (a *IngressReconciler) maybeProvision(ctx context.Context, logger *zap.Suga
continue
}
for _, p := range rule.HTTP.Paths {
// Send a warning if folks use Exact path type - to make
// it easier for us to support Exact path type matching
// in the future if needed.
// https://kubernetes.io/docs/concepts/services-networking/ingress/#path-types
if *p.PathType == networkingv1.PathTypeExact {
msg := "Exact path type strict matching is currently not supported and requests will be routed as for Prefix path type. This behaviour might change in the future."
logger.Warnf(fmt.Sprintf("Unsupported Path type exact for path %s. %s", p.Path, msg))
a.recorder.Eventf(ing, corev1.EventTypeWarning, "UnsupportedPathTypeExact", msg)
}
addIngressBackend(&p.Backend, p.Path)
}
}
if len(web.Handlers) == 0 {
logger.Warn("Ingress contains no valid backends")
a.recorder.Eventf(ing, corev1.EventTypeWarning, "NoValidBackends", "no valid backends")
return nil
}
crl := childResourceLabels(ing.Name, ing.Namespace, "ingress")
var tags []string
if tstr, ok := ing.Annotations[AnnotationTags]; ok {
@@ -226,6 +264,11 @@ func (a *IngressReconciler) maybeProvision(ctx context.Context, logger *zap.Suga
ServeConfig: sc,
Tags: tags,
ChildResourceLabels: crl,
ProxyClass: proxyClass,
}
if val := ing.GetAnnotations()[AnnotationExperimentalForwardClusterTrafficViaL7IngresProxy]; val == "true" {
sts.ForwardClusterTrafficViaL7IngressProxy = true
}
if _, err := a.ssr.Provision(ctx, logger, sts); err != nil {
@@ -267,5 +310,28 @@ func (a *IngressReconciler) maybeProvision(ctx context.Context, logger *zap.Suga
func (a *IngressReconciler) shouldExpose(ing *networkingv1.Ingress) bool {
return ing != nil &&
ing.Spec.IngressClassName != nil &&
*ing.Spec.IngressClassName == "tailscale"
*ing.Spec.IngressClassName == tailscaleIngressClassName
}
// validateIngressClass attempts to validate that 'tailscale' IngressClass
// included in Tailscale installation manifests exists and has not been modified
// to attempt to enable features that we do not support.
func (a *IngressReconciler) validateIngressClass(ctx context.Context) error {
ic := &networkingv1.IngressClass{
ObjectMeta: metav1.ObjectMeta{
Name: tailscaleIngressClassName,
},
}
if err := a.Get(ctx, client.ObjectKeyFromObject(ic), ic); apierrors.IsNotFound(err) {
return errors.New("Tailscale IngressClass not found in cluster. Latest installation manifests include a tailscale IngressClass - please update")
} else if err != nil {
return fmt.Errorf("error retrieving 'tailscale' IngressClass: %w", err)
}
if ic.Spec.Controller != tailscaleIngressControllerName {
return fmt.Errorf("Tailscale Ingress class controller name %s does not match tailscale Ingress controller name %s. Ensure that you are using 'tailscale' IngressClass from latest Tailscale installation manifests", ic.Spec.Controller, tailscaleIngressControllerName)
}
if ic.GetAnnotations()[ingressClassDefaultAnnotation] != "" {
return fmt.Errorf("%s annotation is set on 'tailscale' IngressClass, but Tailscale Ingress controller does not support default Ingress class. Ensure that you are using 'tailscale' IngressClass from latest Tailscale installation manifests", ingressClassDefaultAnnotation)
}
return nil
}

View File

@@ -0,0 +1,270 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
//go:build !plan9
package main
import (
"testing"
"go.uber.org/zap"
appsv1 "k8s.io/api/apps/v1"
corev1 "k8s.io/api/core/v1"
networkingv1 "k8s.io/api/networking/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"sigs.k8s.io/controller-runtime/pkg/client/fake"
"tailscale.com/ipn"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/types/ptr"
"tailscale.com/util/mak"
)
func TestTailscaleIngress(t *testing.T) {
tsIngressClass := &networkingv1.IngressClass{ObjectMeta: metav1.ObjectMeta{Name: "tailscale"}, Spec: networkingv1.IngressClassSpec{Controller: "tailscale.com/ts-ingress"}}
fc := fake.NewFakeClient(tsIngressClass)
ft := &fakeTSClient{}
fakeTsnetServer := &fakeTSNetServer{certDomains: []string{"foo.com"}}
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
ingR := &IngressReconciler{
Client: fc,
ssr: &tailscaleSTSReconciler{
Client: fc,
tsClient: ft,
tsnetServer: fakeTsnetServer,
defaultTags: []string{"tag:k8s"},
operatorNamespace: "operator-ns",
proxyImage: "tailscale/tailscale",
},
logger: zl.Sugar(),
}
// 1. Resources get created for regular Ingress
ing := &networkingv1.Ingress{
TypeMeta: metav1.TypeMeta{Kind: "Ingress", APIVersion: "networking.k8s.io/v1"},
ObjectMeta: metav1.ObjectMeta{
Name: "test",
Namespace: "default",
// The apiserver is supposed to set the UID, but the fake client
// doesn't. So, set it explicitly because other code later depends
// on it being set.
UID: types.UID("1234-UID"),
},
Spec: networkingv1.IngressSpec{
IngressClassName: ptr.To("tailscale"),
DefaultBackend: &networkingv1.IngressBackend{
Service: &networkingv1.IngressServiceBackend{
Name: "test",
Port: networkingv1.ServiceBackendPort{
Number: 8080,
},
},
},
TLS: []networkingv1.IngressTLS{
{Hosts: []string{"default-test"}},
},
},
}
mustCreate(t, fc, ing)
mustCreate(t, fc, &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "test",
Namespace: "default",
},
Spec: corev1.ServiceSpec{
ClusterIP: "1.2.3.4",
Ports: []corev1.ServicePort{{
Port: 8080,
Name: "http"},
},
},
})
expectReconciled(t, ingR, "default", "test")
fullName, shortName := findGenName(t, fc, "default", "test", "ingress")
opts := configOpts{
stsName: shortName,
secretName: fullName,
namespace: "default",
parentType: "ingress",
hostname: "default-test",
}
serveConfig := &ipn.ServeConfig{
TCP: map[uint16]*ipn.TCPPortHandler{443: {HTTPS: true}},
Web: map[ipn.HostPort]*ipn.WebServerConfig{"${TS_CERT_DOMAIN}:443": {Handlers: map[string]*ipn.HTTPHandler{"/": {Proxy: "http://1.2.3.4:8080/"}}}},
}
opts.serveConfig = serveConfig
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "ingress"), nil)
expectEqual(t, fc, expectedSTSUserspace(t, fc, opts), removeHashAnnotation)
// 2. Ingress status gets updated with ingress proxy's MagicDNS name
// once that becomes available.
mustUpdate(t, fc, "operator-ns", opts.secretName, func(secret *corev1.Secret) {
mak.Set(&secret.Data, "device_id", []byte("1234"))
mak.Set(&secret.Data, "device_fqdn", []byte("foo.tailnetxyz.ts.net"))
})
expectReconciled(t, ingR, "default", "test")
ing.Finalizers = append(ing.Finalizers, "tailscale.com/finalizer")
ing.Status.LoadBalancer = networkingv1.IngressLoadBalancerStatus{
Ingress: []networkingv1.IngressLoadBalancerIngress{
{Hostname: "foo.tailnetxyz.ts.net", Ports: []networkingv1.IngressPortStatus{{Port: 443, Protocol: "TCP"}}},
},
}
expectEqual(t, fc, ing, nil)
// 3. Resources get created for Ingress that should allow forwarding
// cluster traffic
mustUpdate(t, fc, "default", "test", func(ing *networkingv1.Ingress) {
mak.Set(&ing.ObjectMeta.Annotations, AnnotationExperimentalForwardClusterTrafficViaL7IngresProxy, "true")
})
opts.shouldEnableForwardingClusterTrafficViaIngress = true
expectReconciled(t, ingR, "default", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 4. Resources get cleaned up when Ingress class is unset
mustUpdate(t, fc, "default", "test", func(ing *networkingv1.Ingress) {
ing.Spec.IngressClassName = ptr.To("nginx")
})
expectReconciled(t, ingR, "default", "test")
expectReconciled(t, ingR, "default", "test") // deleting Ingress STS requires two reconciles
expectMissing[appsv1.StatefulSet](t, fc, "operator-ns", shortName)
expectMissing[corev1.Service](t, fc, "operator-ns", shortName)
expectMissing[corev1.Secret](t, fc, "operator-ns", fullName)
}
func TestTailscaleIngressWithProxyClass(t *testing.T) {
// Setup
pc := &tsapi.ProxyClass{
ObjectMeta: metav1.ObjectMeta{Name: "custom-metadata"},
Spec: tsapi.ProxyClassSpec{StatefulSet: &tsapi.StatefulSet{
Labels: map[string]string{"foo": "bar"},
Annotations: map[string]string{"bar.io/foo": "some-val"},
Pod: &tsapi.Pod{Annotations: map[string]string{"foo.io/bar": "some-val"}}}},
}
tsIngressClass := &networkingv1.IngressClass{ObjectMeta: metav1.ObjectMeta{Name: "tailscale"}, Spec: networkingv1.IngressClassSpec{Controller: "tailscale.com/ts-ingress"}}
fc := fake.NewClientBuilder().
WithScheme(tsapi.GlobalScheme).
WithObjects(pc, tsIngressClass).
WithStatusSubresource(pc).
Build()
ft := &fakeTSClient{}
fakeTsnetServer := &fakeTSNetServer{certDomains: []string{"foo.com"}}
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
ingR := &IngressReconciler{
Client: fc,
ssr: &tailscaleSTSReconciler{
Client: fc,
tsClient: ft,
tsnetServer: fakeTsnetServer,
defaultTags: []string{"tag:k8s"},
operatorNamespace: "operator-ns",
proxyImage: "tailscale/tailscale",
},
logger: zl.Sugar(),
}
// 1. Ingress is created with no ProxyClass specified, default proxy
// resources get configured.
ing := &networkingv1.Ingress{
TypeMeta: metav1.TypeMeta{Kind: "Ingress", APIVersion: "networking.k8s.io/v1"},
ObjectMeta: metav1.ObjectMeta{
Name: "test",
Namespace: "default",
// The apiserver is supposed to set the UID, but the fake client
// doesn't. So, set it explicitly because other code later depends
// on it being set.
UID: types.UID("1234-UID"),
},
Spec: networkingv1.IngressSpec{
IngressClassName: ptr.To("tailscale"),
DefaultBackend: &networkingv1.IngressBackend{
Service: &networkingv1.IngressServiceBackend{
Name: "test",
Port: networkingv1.ServiceBackendPort{
Number: 8080,
},
},
},
TLS: []networkingv1.IngressTLS{
{Hosts: []string{"default-test"}},
},
},
}
mustCreate(t, fc, ing)
mustCreate(t, fc, &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "test",
Namespace: "default",
},
Spec: corev1.ServiceSpec{
ClusterIP: "1.2.3.4",
Ports: []corev1.ServicePort{{
Port: 8080,
Name: "http"},
},
},
})
expectReconciled(t, ingR, "default", "test")
fullName, shortName := findGenName(t, fc, "default", "test", "ingress")
opts := configOpts{
stsName: shortName,
secretName: fullName,
namespace: "default",
parentType: "ingress",
hostname: "default-test",
}
serveConfig := &ipn.ServeConfig{
TCP: map[uint16]*ipn.TCPPortHandler{443: {HTTPS: true}},
Web: map[ipn.HostPort]*ipn.WebServerConfig{"${TS_CERT_DOMAIN}:443": {Handlers: map[string]*ipn.HTTPHandler{"/": {Proxy: "http://1.2.3.4:8080/"}}}},
}
opts.serveConfig = serveConfig
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "ingress"), nil)
expectEqual(t, fc, expectedSTSUserspace(t, fc, opts), removeHashAnnotation)
// 2. Ingress is updated to specify a ProxyClass, ProxyClass is not yet
// ready, so proxy resource configuration does not change.
mustUpdate(t, fc, "default", "test", func(ing *networkingv1.Ingress) {
mak.Set(&ing.ObjectMeta.Labels, LabelProxyClass, "custom-metadata")
})
expectReconciled(t, ingR, "default", "test")
expectEqual(t, fc, expectedSTSUserspace(t, fc, opts), removeHashAnnotation)
// 3. ProxyClass is set to Ready by proxy-class reconciler. Ingress get
// reconciled and configuration from the ProxyClass is applied to the
// created proxy resources.
mustUpdateStatus(t, fc, "", "custom-metadata", func(pc *tsapi.ProxyClass) {
pc.Status = tsapi.ProxyClassStatus{
Conditions: []tsapi.ConnectorCondition{{
Status: metav1.ConditionTrue,
Type: tsapi.ProxyClassready,
ObservedGeneration: pc.Generation,
}}}
})
expectReconciled(t, ingR, "default", "test")
opts.proxyClass = pc.Name
expectEqual(t, fc, expectedSTSUserspace(t, fc, opts), removeHashAnnotation)
// 4. tailscale.com/proxy-class label is removed from the Ingress, the
// Ingress gets reconciled and the custom ProxyClass configuration is
// removed from the proxy resources.
mustUpdate(t, fc, "default", "test", func(ing *networkingv1.Ingress) {
delete(ing.ObjectMeta.Labels, LabelProxyClass)
})
expectReconciled(t, ingR, "default", "test")
opts.proxyClass = ""
expectEqual(t, fc, expectedSTSUserspace(t, fc, opts), removeHashAnnotation)
}

View File

@@ -45,11 +45,14 @@ import (
)
// Generate static manifests for deploying Tailscale operator on Kubernetes from the operator's Helm chart.
//go:generate go run tailscale.com/cmd/k8s-operator/generate
//go:generate go run tailscale.com/cmd/k8s-operator/generate staticmanifests
// Generate Connector CustomResourceDefinition yaml from its Go types.
// Generate Connector and ProxyClass CustomResourceDefinition yamls from their Go types.
//go:generate go run sigs.k8s.io/controller-tools/cmd/controller-gen crd schemapatch:manifests=./deploy/crds output:dir=./deploy/crds paths=../../k8s-operator/apis/...
// Generate CRD docs from the yamls
//go:generate go run fybrik.io/crdoc --resources=./deploy/crds --output=../../k8s-operator/api.md
func main() {
// Required to use our client API. We're fine with the instability since the
// client lives in the same repo as this code.
@@ -62,7 +65,6 @@ func main() {
priorityClassName = defaultEnv("PROXY_PRIORITY_CLASS_NAME", "")
tags = defaultEnv("PROXY_TAGS", "tag:k8s")
tsFirewallMode = defaultEnv("PROXY_FIREWALL_MODE", "")
tsEnableConnector = defaultBool("ENABLE_CONNECTOR", false)
)
var opts []kzap.Opts
@@ -93,7 +95,7 @@ func main() {
maybeLaunchAPIServerProxy(zlog, restConfig, s, mode)
// TODO (irbekrm): gather the reconciler options into an opts struct
// rather than passing a million of them in one by one.
runReconcilers(zlog, s, tsNamespace, restConfig, tsClient, image, priorityClassName, tags, tsFirewallMode, tsEnableConnector)
runReconcilers(zlog, s, tsNamespace, restConfig, tsClient, image, priorityClassName, tags, tsFirewallMode)
}
// initTSNet initializes the tsnet.Server and logs in to Tailscale. It uses the
@@ -201,7 +203,7 @@ waitOnline:
// runReconcilers starts the controller-runtime manager and registers the
// ServiceReconciler. It blocks forever.
func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string, restConfig *rest.Config, tsClient *tailscale.Client, image, priorityClassName, tags, tsFirewallMode string, enableConnector bool) {
func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string, restConfig *rest.Config, tsClient *tailscale.Client, image, priorityClassName, tags, tsFirewallMode string) {
var (
isDefaultLoadBalancer = defaultBool("OPERATOR_DEFAULT_LOAD_BALANCER", false)
)
@@ -216,15 +218,16 @@ func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string
Field: client.InNamespace(tsNamespace).AsSelector(),
}
mgrOpts := manager.Options{
// TODO (irbekrm): stricter filtering what we watch/cache/call
// reconcilers on. c/r by default starts a watch on any
// resources that we GET via the controller manager's client.
Cache: cache.Options{
ByObject: map[client.Object]cache.ByObject{
&corev1.Secret{}: nsFilter,
&appsv1.StatefulSet{}: nsFilter,
},
},
}
if enableConnector {
mgrOpts.Scheme = tsapi.GlobalScheme
Scheme: tsapi.GlobalScheme,
}
mgr, err := manager.New(restConfig, mgrOpts)
if err != nil {
@@ -233,6 +236,9 @@ func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string
svcFilter := handler.EnqueueRequestsFromMapFunc(serviceHandler)
svcChildFilter := handler.EnqueueRequestsFromMapFunc(managedResourceHandlerForType("svc"))
// If a ProxyClassChanges, enqueue all Services labeled with that
// ProxyClass's name.
proxyClassFilterForSvc := handler.EnqueueRequestsFromMapFunc(proxyClassHandlerForSvc(mgr.GetClient(), startlog))
eventRecorder := mgr.GetEventRecorderFor("tailscale-operator")
ssr := &tailscaleSTSReconciler{
@@ -251,6 +257,7 @@ func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string
Watches(&corev1.Service{}, svcFilter).
Watches(&appsv1.StatefulSet{}, svcChildFilter).
Watches(&corev1.Secret{}, svcChildFilter).
Watches(&tsapi.ProxyClass{}, proxyClassFilterForSvc).
Complete(&ServiceReconciler{
ssr: ssr,
Client: mgr.GetClient(),
@@ -259,15 +266,21 @@ func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string
recorder: eventRecorder,
})
if err != nil {
startlog.Fatalf("could not create controller: %v", err)
startlog.Fatalf("could not create service reconciler: %v", err)
}
ingressChildFilter := handler.EnqueueRequestsFromMapFunc(managedResourceHandlerForType("ingress"))
// If a ProxyClassChanges, enqueue all Ingresses labeled with that
// ProxyClass's name.
proxyClassFilterForIngress := handler.EnqueueRequestsFromMapFunc(proxyClassHandlerForIngress(mgr.GetClient(), startlog))
// Enque Ingress if a managed Service or backend Service associated with a tailscale Ingress changes.
svcHandlerForIngress := handler.EnqueueRequestsFromMapFunc(serviceHandlerForIngress(mgr.GetClient(), startlog))
err = builder.
ControllerManagedBy(mgr).
For(&networkingv1.Ingress{}).
Watches(&appsv1.StatefulSet{}, ingressChildFilter).
Watches(&corev1.Secret{}, ingressChildFilter).
Watches(&corev1.Service{}, ingressChildFilter).
Watches(&corev1.Service{}, svcHandlerForIngress).
Watches(&tsapi.ProxyClass{}, proxyClassFilterForIngress).
Complete(&IngressReconciler{
ssr: ssr,
recorder: eventRecorder,
@@ -275,25 +288,38 @@ func runReconcilers(zlog *zap.SugaredLogger, s *tsnet.Server, tsNamespace string
logger: zlog.Named("ingress-reconciler"),
})
if err != nil {
startlog.Fatalf("could not create controller: %v", err)
startlog.Fatalf("could not create ingress reconciler: %v", err)
}
if enableConnector {
connectorFilter := handler.EnqueueRequestsFromMapFunc(managedResourceHandlerForType("subnetrouter"))
err = builder.ControllerManagedBy(mgr).
For(&tsapi.Connector{}).
Watches(&appsv1.StatefulSet{}, connectorFilter).
Watches(&corev1.Secret{}, connectorFilter).
Complete(&ConnectorReconciler{
ssr: ssr,
recorder: eventRecorder,
Client: mgr.GetClient(),
logger: zlog.Named("connector-reconciler"),
clock: tstime.DefaultClock{},
})
if err != nil {
startlog.Fatal("could not create connector reconciler: %v", err)
}
connectorFilter := handler.EnqueueRequestsFromMapFunc(managedResourceHandlerForType("connector"))
// If a ProxyClassChanges, enqueue all Connectors that have
// .spec.proxyClass set to the name of this ProxyClass.
proxyClassFilterForConnector := handler.EnqueueRequestsFromMapFunc(proxyClassHandlerForConnector(mgr.GetClient(), startlog))
err = builder.ControllerManagedBy(mgr).
For(&tsapi.Connector{}).
Watches(&appsv1.StatefulSet{}, connectorFilter).
Watches(&corev1.Secret{}, connectorFilter).
Watches(&tsapi.ProxyClass{}, proxyClassFilterForConnector).
Complete(&ConnectorReconciler{
ssr: ssr,
recorder: eventRecorder,
Client: mgr.GetClient(),
logger: zlog.Named("connector-reconciler"),
clock: tstime.DefaultClock{},
})
if err != nil {
startlog.Fatal("could not create connector reconciler: %v", err)
}
err = builder.ControllerManagedBy(mgr).
For(&tsapi.ProxyClass{}).
Complete(&ProxyClassReconciler{
Client: mgr.GetClient(),
recorder: eventRecorder,
logger: zlog.Named("proxyclass-reconciler"),
clock: tstime.DefaultClock{},
})
if err != nil {
startlog.Fatal("could not create proxyclass reconciler: %v", err)
}
startlog.Infof("Startup complete, operator running, version: %s", version.Long())
if err := mgr.Start(signals.SetupSignalHandler()); err != nil {
@@ -323,6 +349,7 @@ func parentFromObjectLabels(o client.Object) types.NamespacedName {
Name: ls[LabelParentName],
}
}
func managedResourceHandlerForType(typ string) handler.MapFunc {
return func(_ context.Context, o client.Object) []reconcile.Request {
if !isManagedByType(o, typ) {
@@ -332,14 +359,115 @@ func managedResourceHandlerForType(typ string) handler.MapFunc {
{NamespacedName: parentFromObjectLabels(o)},
}
}
}
// proxyClassHandlerForSvc returns a handler that, for a given ProxyClass,
// returns a list of reconcile requests for all Services labeled with
// tailscale.com/proxy-class: <proxy class name>.
func proxyClassHandlerForSvc(cl client.Client, logger *zap.SugaredLogger) handler.MapFunc {
return func(ctx context.Context, o client.Object) []reconcile.Request {
svcList := new(corev1.ServiceList)
labels := map[string]string{
LabelProxyClass: o.GetName(),
}
if err := cl.List(ctx, svcList, client.MatchingLabels(labels)); err != nil {
logger.Debugf("error listing Services for ProxyClass: %v", err)
return nil
}
reqs := make([]reconcile.Request, 0)
for _, svc := range svcList.Items {
reqs = append(reqs, reconcile.Request{NamespacedName: client.ObjectKeyFromObject(&svc)})
}
return reqs
}
}
// proxyClassHandlerForIngress returns a handler that, for a given ProxyClass,
// returns a list of reconcile requests for all Ingresses labeled with
// tailscale.com/proxy-class: <proxy class name>.
func proxyClassHandlerForIngress(cl client.Client, logger *zap.SugaredLogger) handler.MapFunc {
return func(ctx context.Context, o client.Object) []reconcile.Request {
ingList := new(networkingv1.IngressList)
labels := map[string]string{
LabelProxyClass: o.GetName(),
}
if err := cl.List(ctx, ingList, client.MatchingLabels(labels)); err != nil {
logger.Debugf("error listing Ingresses for ProxyClass: %v", err)
return nil
}
reqs := make([]reconcile.Request, 0)
for _, ing := range ingList.Items {
reqs = append(reqs, reconcile.Request{NamespacedName: client.ObjectKeyFromObject(&ing)})
}
return reqs
}
}
// proxyClassHandlerForConnector returns a handler that, for a given ProxyClass,
// returns a list of reconcile requests for all Connectors that have
// .spec.proxyClass set.
func proxyClassHandlerForConnector(cl client.Client, logger *zap.SugaredLogger) handler.MapFunc {
return func(ctx context.Context, o client.Object) []reconcile.Request {
connList := new(tsapi.ConnectorList)
if err := cl.List(ctx, connList); err != nil {
logger.Debugf("error listing Connectors for ProxyClass: %v", err)
return nil
}
reqs := make([]reconcile.Request, 0)
proxyClassName := o.GetName()
for _, conn := range connList.Items {
if conn.Spec.ProxyClass == proxyClassName {
reqs = append(reqs, reconcile.Request{NamespacedName: client.ObjectKeyFromObject(&conn)})
}
}
return reqs
}
}
// serviceHandlerForIngress returns a handler for Service events for ingress
// reconciler that ensures that if the Service associated with an event is of
// interest to the reconciler, the associated Ingress(es) gets be reconciled.
// The Services of interest are backend Services for tailscale Ingress and
// managed Services for an StatefulSet for a proxy configured for tailscale
// Ingress
func serviceHandlerForIngress(cl client.Client, logger *zap.SugaredLogger) handler.MapFunc {
return func(ctx context.Context, o client.Object) []reconcile.Request {
if isManagedByType(o, "ingress") {
ingName := parentFromObjectLabels(o)
return []reconcile.Request{{NamespacedName: ingName}}
}
ingList := networkingv1.IngressList{}
if err := cl.List(ctx, &ingList, client.InNamespace(o.GetNamespace())); err != nil {
logger.Debugf("error listing Ingresses: %v", err)
return nil
}
reqs := make([]reconcile.Request, 0)
for _, ing := range ingList.Items {
if ing.Spec.IngressClassName == nil || *ing.Spec.IngressClassName != tailscaleIngressClassName {
return nil
}
if ing.Spec.DefaultBackend != nil && ing.Spec.DefaultBackend.Service != nil && ing.Spec.DefaultBackend.Service.Name == o.GetName() {
reqs = append(reqs, reconcile.Request{NamespacedName: client.ObjectKeyFromObject(&ing)})
}
for _, rule := range ing.Spec.Rules {
if rule.HTTP == nil {
continue
}
for _, path := range rule.HTTP.Paths {
if path.Backend.Service != nil && path.Backend.Service.Name == o.GetName() {
reqs = append(reqs, reconcile.Request{NamespacedName: client.ObjectKeyFromObject(&ing)})
}
}
}
}
return reqs
}
}
func serviceHandler(_ context.Context, o client.Object) []reconcile.Request {
if isManagedByType(o, "svc") {
// If this is a Service managed by a Service we want to enqueue its parent
return []reconcile.Request{{NamespacedName: parentFromObjectLabels(o)}}
}
if isManagedResource(o) {
// If this is a Servce managed by a resource that is not a Service, we leave it alone
@@ -354,7 +482,6 @@ func serviceHandler(_ context.Context, o client.Object) []reconcile.Request {
},
},
}
}
// isMagicDNSName reports whether name is a full tailnet node FQDN (with or

View File

@@ -6,16 +6,22 @@
package main
import (
"context"
"fmt"
"testing"
"github.com/google/go-cmp/cmp"
"go.uber.org/zap"
appsv1 "k8s.io/api/apps/v1"
corev1 "k8s.io/api/core/v1"
networkingv1 "k8s.io/api/networking/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"sigs.k8s.io/controller-runtime/pkg/client/fake"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/types/ptr"
"tailscale.com/util/mak"
)
func TestLoadBalancerClass(t *testing.T) {
@@ -67,9 +73,9 @@ func TestLoadBalancerClass(t *testing.T) {
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSecret(t, opts))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(opts))
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// Normally the Tailscale proxy pod would come up here and write its info
// into the secret. Simulate that, then verify reconcile again and verify
@@ -112,7 +118,7 @@ func TestLoadBalancerClass(t *testing.T) {
},
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
// Turn the service back into a ClusterIP service, which should make the
// operator clean up.
@@ -151,7 +157,7 @@ func TestLoadBalancerClass(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
}
func TestTailnetTargetFQDNAnnotation(t *testing.T) {
@@ -208,9 +214,9 @@ func TestTailnetTargetFQDNAnnotation(t *testing.T) {
hostname: "default-test",
}
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
want := &corev1.Service{
TypeMeta: metav1.TypeMeta{
Kind: "Service",
@@ -231,10 +237,10 @@ func TestTailnetTargetFQDNAnnotation(t *testing.T) {
Selector: nil,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, want, nil)
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
// Change the tailscale-target-fqdn annotation which should update the
// StatefulSet
@@ -318,9 +324,9 @@ func TestTailnetTargetIPAnnotation(t *testing.T) {
hostname: "default-test",
}
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
want := &corev1.Service{
TypeMeta: metav1.TypeMeta{
Kind: "Service",
@@ -341,10 +347,10 @@ func TestTailnetTargetIPAnnotation(t *testing.T) {
Selector: nil,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, want, nil)
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
// Change the tailscale-target-ip annotation which should update the
// StatefulSet
@@ -425,9 +431,9 @@ func TestAnnotations(t *testing.T) {
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
want := &corev1.Service{
TypeMeta: metav1.TypeMeta{
Kind: "Service",
@@ -447,7 +453,7 @@ func TestAnnotations(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
// Turn the service back into a ClusterIP service, which should make the
// operator clean up.
@@ -479,7 +485,7 @@ func TestAnnotations(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
}
func TestAnnotationIntoLB(t *testing.T) {
@@ -533,9 +539,9 @@ func TestAnnotationIntoLB(t *testing.T) {
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
// Normally the Tailscale proxy pod would come up here and write its info
// into the secret. Simulate that, since it would have normally happened at
@@ -568,7 +574,7 @@ func TestAnnotationIntoLB(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
// Remove Tailscale's annotation, and at the same time convert the service
// into a tailscale LoadBalancer.
@@ -579,8 +585,8 @@ func TestAnnotationIntoLB(t *testing.T) {
})
expectReconciled(t, sr, "default", "test")
// None of the proxy machinery should have changed...
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
// ... but the service should have a LoadBalancer status.
want = &corev1.Service{
@@ -612,7 +618,7 @@ func TestAnnotationIntoLB(t *testing.T) {
},
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
}
func TestLBIntoAnnotation(t *testing.T) {
@@ -664,9 +670,9 @@ func TestLBIntoAnnotation(t *testing.T) {
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
// Normally the Tailscale proxy pod would come up here and write its info
// into the secret. Simulate that, then verify reconcile again and verify
@@ -709,7 +715,7 @@ func TestLBIntoAnnotation(t *testing.T) {
},
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
// Turn the service back into a ClusterIP service, but also add the
// tailscale annotation.
@@ -728,8 +734,8 @@ func TestLBIntoAnnotation(t *testing.T) {
})
expectReconciled(t, sr, "default", "test")
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
want = &corev1.Service{
TypeMeta: metav1.TypeMeta{
@@ -750,7 +756,7 @@ func TestLBIntoAnnotation(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
}
func TestCustomHostname(t *testing.T) {
@@ -805,9 +811,9 @@ func TestCustomHostname(t *testing.T) {
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSecret(t, o))
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSecret(t, o), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
want := &corev1.Service{
TypeMeta: metav1.TypeMeta{
Kind: "Service",
@@ -828,7 +834,7 @@ func TestCustomHostname(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
// Turn the service back into a ClusterIP service, which should make the
// operator clean up.
@@ -863,7 +869,7 @@ func TestCustomHostname(t *testing.T) {
Type: corev1.ServiceTypeClusterIP,
},
}
expectEqual(t, fc, want)
expectEqual(t, fc, want, nil)
}
func TestCustomPriorityClassName(t *testing.T) {
@@ -920,7 +926,104 @@ func TestCustomPriorityClassName(t *testing.T) {
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
}
func TestProxyClassForService(t *testing.T) {
// Setup
pc := &tsapi.ProxyClass{
ObjectMeta: metav1.ObjectMeta{Name: "custom-metadata"},
Spec: tsapi.ProxyClassSpec{StatefulSet: &tsapi.StatefulSet{
Labels: map[string]string{"foo": "bar"},
Annotations: map[string]string{"bar.io/foo": "some-val"},
Pod: &tsapi.Pod{Annotations: map[string]string{"foo.io/bar": "some-val"}}}},
}
fc := fake.NewClientBuilder().
WithScheme(tsapi.GlobalScheme).
WithObjects(pc).
WithStatusSubresource(pc).
Build()
ft := &fakeTSClient{}
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
sr := &ServiceReconciler{
Client: fc,
ssr: &tailscaleSTSReconciler{
Client: fc,
tsClient: ft,
defaultTags: []string{"tag:k8s"},
operatorNamespace: "operator-ns",
proxyImage: "tailscale/tailscale",
},
logger: zl.Sugar(),
}
// 1. A new tailscale LoadBalancer Service is created without any
// ProxyClass. Resources get created for it as usual.
mustCreate(t, fc, &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "test",
Namespace: "default",
// The apiserver is supposed to set the UID, but the fake client
// doesn't. So, set it explicitly because other code later depends
// on it being set.
UID: types.UID("1234-UID"),
},
Spec: corev1.ServiceSpec{
ClusterIP: "10.20.30.40",
Type: corev1.ServiceTypeLoadBalancer,
LoadBalancerClass: ptr.To("tailscale"),
},
})
expectReconciled(t, sr, "default", "test")
fullName, shortName := findGenName(t, fc, "default", "test", "svc")
opts := configOpts{
stsName: shortName,
secretName: fullName,
namespace: "default",
parentType: "svc",
hostname: "default-test",
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSecret(t, opts), nil)
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 2. The Service gets updated with tailscale.com/proxy-class label
// pointing at the 'custom-metadata' ProxyClass. The ProxyClass is not
// yet ready, so no changes are actually applied to the proxy resources.
mustUpdate(t, fc, "default", "test", func(svc *corev1.Service) {
mak.Set(&svc.Labels, LabelProxyClass, "custom-metadata")
})
expectReconciled(t, sr, "default", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 3. ProxyClass is set to Ready, the Service gets reconciled by the
// services-reconciler and the customization from the ProxyClass is
// applied to the proxy resources.
mustUpdateStatus(t, fc, "", "custom-metadata", func(pc *tsapi.ProxyClass) {
pc.Status = tsapi.ProxyClassStatus{
Conditions: []tsapi.ConnectorCondition{{
Status: metav1.ConditionTrue,
Type: tsapi.ProxyClassready,
ObservedGeneration: pc.Generation,
}}}
})
opts.proxyClass = pc.Name
expectReconciled(t, sr, "default", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
// 4. tailscale.com/proxy-class label is removed from the Service, the
// configuration from the ProxyClass is removed from the cluster
// resources.
mustUpdate(t, fc, "default", "test", func(svc *corev1.Service) {
delete(svc.Labels, LabelProxyClass)
})
opts.proxyClass = ""
expectReconciled(t, sr, "default", "test")
expectEqual(t, fc, expectedSTS(t, fc, opts), removeHashAnnotation)
}
func TestDefaultLoadBalancer(t *testing.T) {
@@ -964,7 +1067,7 @@ func TestDefaultLoadBalancer(t *testing.T) {
fullName, shortName := findGenName(t, fc, "default", "test", "svc")
expectEqual(t, fc, expectedHeadlessService(shortName))
expectEqual(t, fc, expectedHeadlessService(shortName, "svc"), nil)
o := configOpts{
stsName: shortName,
secretName: fullName,
@@ -973,7 +1076,8 @@ func TestDefaultLoadBalancer(t *testing.T) {
hostname: "default-test",
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
}
func TestProxyFirewallMode(t *testing.T) {
@@ -1026,8 +1130,69 @@ func TestProxyFirewallMode(t *testing.T) {
firewallMode: "nftables",
clusterTargetIP: "10.20.30.40",
}
expectEqual(t, fc, expectedSTS(o))
expectEqual(t, fc, expectedSTS(t, fc, o), removeHashAnnotation)
}
func TestTailscaledConfigfileHash(t *testing.T) {
fc := fake.NewFakeClient()
ft := &fakeTSClient{}
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
sr := &ServiceReconciler{
Client: fc,
ssr: &tailscaleSTSReconciler{
Client: fc,
tsClient: ft,
defaultTags: []string{"tag:k8s"},
operatorNamespace: "operator-ns",
proxyImage: "tailscale/tailscale",
},
logger: zl.Sugar(),
isDefaultLoadBalancer: true,
}
// Create a service that we should manage, and check that the initial round
// of objects looks right.
mustCreate(t, fc, &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "test",
Namespace: "default",
// The apiserver is supposed to set the UID, but the fake client
// doesn't. So, set it explicitly because other code later depends
// on it being set.
UID: types.UID("1234-UID"),
},
Spec: corev1.ServiceSpec{
ClusterIP: "10.20.30.40",
Type: corev1.ServiceTypeLoadBalancer,
},
})
expectReconciled(t, sr, "default", "test")
fullName, shortName := findGenName(t, fc, "default", "test", "svc")
o := configOpts{
stsName: shortName,
secretName: fullName,
namespace: "default",
parentType: "svc",
hostname: "default-test",
clusterTargetIP: "10.20.30.40",
confFileHash: "705e5ffd0bd5326237efdf542c850a65a54101284d5daa30775420fcc64d89c1",
}
expectEqual(t, fc, expectedSTS(t, fc, o), nil)
// 2. Hostname gets changed, configfile is updated and a new hash value
// is produced.
mustUpdate(t, fc, "default", "test", func(svc *corev1.Service) {
mak.Set(&svc.Annotations, AnnotationHostname, "another-test")
})
o.hostname = "another-test"
o.confFileHash = "1a087f887825d2b75d3673c7c2b0131f8ec1f0b1cb761d33e236dd28350dfe23"
expectReconciled(t, sr, "default", "test")
expectEqual(t, fc, expectedSTS(t, fc, o), nil)
}
func Test_isMagicDNSName(t *testing.T) {
@@ -1056,3 +1221,134 @@ func Test_isMagicDNSName(t *testing.T) {
})
}
}
func Test_serviceHandlerForIngress(t *testing.T) {
fc := fake.NewFakeClient()
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
// 1. An event on a headless Service for a tailscale Ingress results in
// the Ingress being reconciled.
mustCreate(t, fc, &networkingv1.Ingress{
ObjectMeta: metav1.ObjectMeta{
Name: "ing-1",
Namespace: "ns-1",
},
Spec: networkingv1.IngressSpec{IngressClassName: ptr.To(tailscaleIngressClassName)},
})
svc1 := &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "headless-1",
Namespace: "tailscale",
Labels: map[string]string{
LabelManaged: "true",
LabelParentName: "ing-1",
LabelParentNamespace: "ns-1",
LabelParentType: "ingress",
},
},
}
mustCreate(t, fc, svc1)
wantReqs := []reconcile.Request{{NamespacedName: types.NamespacedName{Namespace: "ns-1", Name: "ing-1"}}}
gotReqs := serviceHandlerForIngress(fc, zl.Sugar())(context.Background(), svc1)
if diff := cmp.Diff(gotReqs, wantReqs); diff != "" {
t.Fatalf("unexpected reconcile requests (-got +want):\n%s", diff)
}
// 2. An event on a Service that is the default backend for a tailscale
// Ingress results in the Ingress being reconciled.
mustCreate(t, fc, &networkingv1.Ingress{
ObjectMeta: metav1.ObjectMeta{
Name: "ing-2",
Namespace: "ns-2",
},
Spec: networkingv1.IngressSpec{
DefaultBackend: &networkingv1.IngressBackend{
Service: &networkingv1.IngressServiceBackend{Name: "def-backend"},
},
IngressClassName: ptr.To(tailscaleIngressClassName),
},
})
backendSvc := &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "def-backend",
Namespace: "ns-2",
},
}
mustCreate(t, fc, backendSvc)
wantReqs = []reconcile.Request{{NamespacedName: types.NamespacedName{Namespace: "ns-2", Name: "ing-2"}}}
gotReqs = serviceHandlerForIngress(fc, zl.Sugar())(context.Background(), backendSvc)
if diff := cmp.Diff(gotReqs, wantReqs); diff != "" {
t.Fatalf("unexpected reconcile requests (-got +want):\n%s", diff)
}
// 3. An event on a Service that is one of the non-default backends for
// a tailscale Ingress results in the Ingress being reconciled.
mustCreate(t, fc, &networkingv1.Ingress{
ObjectMeta: metav1.ObjectMeta{
Name: "ing-3",
Namespace: "ns-3",
},
Spec: networkingv1.IngressSpec{
IngressClassName: ptr.To(tailscaleIngressClassName),
Rules: []networkingv1.IngressRule{{IngressRuleValue: networkingv1.IngressRuleValue{HTTP: &networkingv1.HTTPIngressRuleValue{
Paths: []networkingv1.HTTPIngressPath{
{Backend: networkingv1.IngressBackend{Service: &networkingv1.IngressServiceBackend{Name: "backend"}}}},
}}}},
},
})
backendSvc2 := &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "backend",
Namespace: "ns-3",
},
}
mustCreate(t, fc, backendSvc2)
wantReqs = []reconcile.Request{{NamespacedName: types.NamespacedName{Namespace: "ns-3", Name: "ing-3"}}}
gotReqs = serviceHandlerForIngress(fc, zl.Sugar())(context.Background(), backendSvc2)
if diff := cmp.Diff(gotReqs, wantReqs); diff != "" {
t.Fatalf("unexpected reconcile requests (-got +want):\n%s", diff)
}
// 4. An event on a Service that is a backend for an Ingress that is not
// tailscale Ingress does not result in an Ingress reconcile.
mustCreate(t, fc, &networkingv1.Ingress{
ObjectMeta: metav1.ObjectMeta{
Name: "ing-4",
Namespace: "ns-4",
},
Spec: networkingv1.IngressSpec{
Rules: []networkingv1.IngressRule{{IngressRuleValue: networkingv1.IngressRuleValue{HTTP: &networkingv1.HTTPIngressRuleValue{
Paths: []networkingv1.HTTPIngressPath{
{Backend: networkingv1.IngressBackend{Service: &networkingv1.IngressServiceBackend{Name: "non-ts-backend"}}}},
}}}},
},
})
nonTSBackend := &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "non-ts-backend",
Namespace: "ns-4",
},
}
mustCreate(t, fc, nonTSBackend)
gotReqs = serviceHandlerForIngress(fc, zl.Sugar())(context.Background(), nonTSBackend)
if len(gotReqs) > 0 {
t.Errorf("unexpected reconcile request for a Service that does not belong to a Tailscale Ingress: %#+v\n", gotReqs)
}
// 5. An event on a Service not related to any Ingress does not result
// in an Ingress reconcile.
someSvc := &corev1.Service{
ObjectMeta: metav1.ObjectMeta{
Name: "some-svc",
Namespace: "ns-4",
},
}
mustCreate(t, fc, someSvc)
gotReqs = serviceHandlerForIngress(fc, zl.Sugar())(context.Background(), someSvc)
if len(gotReqs) > 0 {
t.Errorf("unexpected reconcile request for a Service that does not belong to any Ingress: %#+v\n", gotReqs)
}
}

View File

@@ -6,7 +6,6 @@
package main
import (
"context"
"crypto/tls"
"fmt"
"log"
@@ -24,22 +23,11 @@ import (
"tailscale.com/tailcfg"
"tailscale.com/tsnet"
"tailscale.com/util/clientmetric"
"tailscale.com/util/ctxkey"
"tailscale.com/util/set"
)
type whoIsKey struct{}
// whoIsFromRequest returns the WhoIsResponse previously stashed by a call to
// addWhoIsToRequest.
func whoIsFromRequest(r *http.Request) *apitype.WhoIsResponse {
return r.Context().Value(whoIsKey{}).(*apitype.WhoIsResponse)
}
// addWhoIsToRequest stashes who in r's context, retrievable by a call to
// whoIsFromRequest.
func addWhoIsToRequest(r *http.Request, who *apitype.WhoIsResponse) *http.Request {
return r.WithContext(context.WithValue(r.Context(), whoIsKey{}, who))
}
var whoIsKey = ctxkey.New("", (*apitype.WhoIsResponse)(nil))
var counterNumRequestsProxied = clientmetric.NewCounter("k8s_auth_proxy_requests_proxied")
@@ -127,7 +115,7 @@ func (h *apiserverProxy) ServeHTTP(w http.ResponseWriter, r *http.Request) {
return
}
counterNumRequestsProxied.Add(1)
h.rp.ServeHTTP(w, addWhoIsToRequest(r, who))
h.rp.ServeHTTP(w, r.WithContext(whoIsKey.WithValue(r.Context(), who)))
}
// runAPIServerProxy runs an HTTP server that authenticates requests using the
@@ -240,7 +228,7 @@ type impersonateRule struct {
// in the context by the apiserverProxy.
func addImpersonationHeaders(r *http.Request, log *zap.SugaredLogger) error {
log = log.With("remote", r.RemoteAddr)
who := whoIsFromRequest(r)
who := whoIsKey.Value(r.Context())
rules, err := tailcfg.UnmarshalCapJSON[capRule](who.CapMap, capabilityName)
if len(rules) == 0 && err == nil {
// Try the old capability name for backwards compatibility.

View File

@@ -95,7 +95,7 @@ func TestImpersonationHeaders(t *testing.T) {
for _, tc := range tests {
r := must.Get(http.NewRequest("GET", "https://op.ts.net/api/foo", nil))
r = addWhoIsToRequest(r, &apitype.WhoIsResponse{
r = r.WithContext(whoIsKey.WithValue(r.Context(), &apitype.WhoIsResponse{
Node: &tailcfg.Node{
Name: "node.ts.net",
Tags: tc.tags,
@@ -104,7 +104,7 @@ func TestImpersonationHeaders(t *testing.T) {
LoginName: tc.emailish,
},
CapMap: tc.capMap,
})
}))
addImpersonationHeaders(r, zl.Sugar())
if d := cmp.Diff(tc.wantHeaders, r.Header); d != "" {

View File

@@ -0,0 +1,108 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
//go:build !plan9
// tailscale-operator provides a way to expose services running in a Kubernetes
// cluster to your Tailnet.
package main
import (
"context"
"fmt"
"go.uber.org/zap"
corev1 "k8s.io/api/core/v1"
apiequality "k8s.io/apimachinery/pkg/api/equality"
apierrors "k8s.io/apimachinery/pkg/api/errors"
apivalidation "k8s.io/apimachinery/pkg/api/validation"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
metavalidation "k8s.io/apimachinery/pkg/apis/meta/v1/validation"
"k8s.io/apimachinery/pkg/util/validation/field"
"k8s.io/client-go/tools/record"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
tsoperator "tailscale.com/k8s-operator"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/tstime"
)
const (
reasonProxyClassInvalid = "ProxyClassInvalid"
reasonProxyClassValid = "ProxyClassValid"
messageProxyClassInvalid = "ProxyClass is not valid: %v"
)
type ProxyClassReconciler struct {
client.Client
recorder record.EventRecorder
logger *zap.SugaredLogger
clock tstime.Clock
}
func (pcr *ProxyClassReconciler) Reconcile(ctx context.Context, req reconcile.Request) (res reconcile.Result, err error) {
logger := pcr.logger.With("ProxyClass", req.Name)
logger.Debugf("starting reconcile")
defer logger.Debugf("reconcile finished")
pc := new(tsapi.ProxyClass)
err = pcr.Get(ctx, req.NamespacedName, pc)
if apierrors.IsNotFound(err) {
logger.Debugf("ProxyClass not found, assuming it was deleted")
return reconcile.Result{}, nil
} else if err != nil {
return reconcile.Result{}, fmt.Errorf("failed to get tailscale.com ProxyClass: %w", err)
}
if !pc.DeletionTimestamp.IsZero() {
logger.Debugf("ProxyClass is being deleted, do nothing")
return reconcile.Result{}, nil
}
oldPCStatus := pc.Status.DeepCopy()
if errs := pcr.validate(pc); errs != nil {
msg := fmt.Sprintf(messageProxyClassInvalid, errs.ToAggregate().Error())
pcr.recorder.Event(pc, corev1.EventTypeWarning, reasonProxyClassInvalid, msg)
tsoperator.SetProxyClassCondition(pc, tsapi.ProxyClassready, metav1.ConditionFalse, reasonProxyClassInvalid, msg, pc.Generation, pcr.clock, logger)
} else {
tsoperator.SetProxyClassCondition(pc, tsapi.ProxyClassready, metav1.ConditionTrue, reasonProxyClassValid, reasonProxyClassValid, pc.Generation, pcr.clock, logger)
}
if !apiequality.Semantic.DeepEqual(oldPCStatus, pc.Status) {
if err := pcr.Client.Status().Update(ctx, pc); err != nil {
logger.Errorf("error updating ProxyClass status: %v", err)
return reconcile.Result{}, err
}
}
return reconcile.Result{}, nil
}
func (a *ProxyClassReconciler) validate(pc *tsapi.ProxyClass) (violations field.ErrorList) {
if sts := pc.Spec.StatefulSet; sts != nil {
if len(sts.Labels) > 0 {
if errs := metavalidation.ValidateLabels(sts.Labels, field.NewPath(".spec.statefulSet.labels")); errs != nil {
violations = append(violations, errs...)
}
}
if len(sts.Annotations) > 0 {
if errs := apivalidation.ValidateAnnotations(sts.Annotations, field.NewPath(".spec.statefulSet.annotations")); errs != nil {
violations = append(violations, errs...)
}
}
if pod := sts.Pod; pod != nil {
if len(pod.Labels) > 0 {
if errs := metavalidation.ValidateLabels(pod.Labels, field.NewPath(".spec.statefulSet.pod.labels")); errs != nil {
violations = append(violations, errs...)
}
}
if len(pod.Annotations) > 0 {
if errs := apivalidation.ValidateAnnotations(pod.Annotations, field.NewPath(".spec.statefulSet.pod.annotations")); errs != nil {
violations = append(violations, errs...)
}
}
}
}
// We do not validate embedded fields (security context, resource
// requirements etc) as we inherit upstream validation for those fields.
// Invalid values would get rejected by upstream validations at apply
// time.
return violations
}

View File

@@ -0,0 +1,83 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
//go:build !plan9
// tailscale-operator provides a way to expose services running in a Kubernetes
// cluster to your Tailnet.
package main
import (
"testing"
"time"
"go.uber.org/zap"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"k8s.io/client-go/tools/record"
"sigs.k8s.io/controller-runtime/pkg/client/fake"
tsoperator "tailscale.com/k8s-operator"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/tstest"
)
func TestProxyClass(t *testing.T) {
pc := &tsapi.ProxyClass{
TypeMeta: metav1.TypeMeta{Kind: "ProxyClass", APIVersion: "tailscale.com/v1alpha1"},
ObjectMeta: metav1.ObjectMeta{
Name: "test",
// The apiserver is supposed to set the UID, but the fake client
// doesn't. So, set it explicitly because other code later depends
// on it being set.
UID: types.UID("1234-UID"),
},
Spec: tsapi.ProxyClassSpec{
StatefulSet: &tsapi.StatefulSet{
Labels: map[string]string{"foo": "bar", "xyz1234": "abc567"},
Annotations: map[string]string{"foo.io/bar": "{'key': 'val1232'}"},
Pod: &tsapi.Pod{
Labels: map[string]string{"foo": "bar", "xyz1234": "abc567"},
Annotations: map[string]string{"foo.io/bar": "{'key': 'val1232'}"},
},
},
},
}
fc := fake.NewClientBuilder().
WithScheme(tsapi.GlobalScheme).
WithObjects(pc).
WithStatusSubresource(pc).
Build()
zl, err := zap.NewDevelopment()
if err != nil {
t.Fatal(err)
}
cl := tstest.NewClock(tstest.ClockOpts{})
pcr := &ProxyClassReconciler{
Client: fc,
logger: zl.Sugar(),
clock: cl,
recorder: record.NewFakeRecorder(1),
}
expectReconciled(t, pcr, "", "test")
// 1. A valid ProxyClass resource gets its status updated to Ready.
pc.Status.Conditions = append(pc.Status.Conditions, tsapi.ConnectorCondition{
Type: tsapi.ProxyClassready,
Status: metav1.ConditionTrue,
Reason: reasonProxyClassValid,
Message: reasonProxyClassValid,
LastTransitionTime: &metav1.Time{Time: cl.Now().Truncate(time.Second)},
})
expectEqual(t, fc, pc, nil)
// 2. An invalid ProxyClass resource gets its status updated to Invalid.
pc.Spec.StatefulSet.Labels["foo"] = "?!someVal"
mustUpdate(t, fc, "", "test", func(proxyClass *tsapi.ProxyClass) {
proxyClass.Spec.StatefulSet.Labels = pc.Spec.StatefulSet.Labels
})
expectReconciled(t, pcr, "", "test")
msg := `ProxyClass is not valid: .spec.statefulSet.labels: Invalid value: "?!someVal": a valid label must be an empty string or consist of alphanumeric characters, '-', '_' or '.', and must start and end with an alphanumeric character (e.g. 'MyValue', or 'my_value', or '12345', regex used for validation is '(([A-Za-z0-9][-A-Za-z0-9_.]*)?[A-Za-z0-9])?')`
tsoperator.SetProxyClassCondition(pc, tsapi.ProxyClassready, metav1.ConditionFalse, reasonProxyClassInvalid, msg, 0, cl, zl.Sugar())
expectEqual(t, fc, pc, nil)
}

View File

@@ -14,6 +14,7 @@ import (
"fmt"
"net/http"
"os"
"slices"
"strings"
"go.uber.org/zap"
@@ -22,25 +23,37 @@ import (
apierrors "k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/apis/meta/v1/unstructured"
"k8s.io/apimachinery/pkg/types"
"k8s.io/apiserver/pkg/storage/names"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/yaml"
"tailscale.com/client/tailscale"
"tailscale.com/ipn"
tsoperator "tailscale.com/k8s-operator"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/net/netutil"
"tailscale.com/tailcfg"
"tailscale.com/tsnet"
"tailscale.com/types/opt"
"tailscale.com/types/ptr"
"tailscale.com/util/dnsname"
"tailscale.com/util/mak"
)
const (
// Labels that the operator sets on StatefulSets and Pods. If you add a
// new label here, do also add it to tailscaleManagedLabels var to
// ensure that it does not get overwritten by ProxyClass configuration.
LabelManaged = "tailscale.com/managed"
LabelParentType = "tailscale.com/parent-resource-type"
LabelParentName = "tailscale.com/parent-resource"
LabelParentNamespace = "tailscale.com/parent-resource-ns"
// LabelProxyClass can be set by users on Connectors, tailscale
// Ingresses and Services that define cluster ingress or cluster egress,
// to specify that configuration in this ProxyClass should be applied to
// resources created for the Connector, Ingress or Service.
LabelProxyClass = "tailscale.com/proxy-class"
FinalizerName = "tailscale.com/finalizer"
// Annotations settable by users on services.
@@ -55,10 +68,25 @@ const (
// Annotations settable by users on ingresses.
AnnotationFunnel = "tailscale.com/funnel"
// If set to true, set up iptables/nftables rules in the proxy forward
// cluster traffic to the tailnet IP of that proxy. This can only be set
// on an Ingress. This is useful in cases where a cluster target needs
// to be able to reach a cluster workload exposed to tailnet via Ingress
// using the same hostname as a tailnet workload (in this case, the
// MagicDNS name of the ingress proxy). This annotation is experimental.
// If it is set to true, the proxy set up for Ingress, will run
// tailscale in non-userspace, with NET_ADMIN cap for tailscale
// container and will also run a privileged init container that enables
// forwarding.
// Eventually this behaviour might become the default.
AnnotationExperimentalForwardClusterTrafficViaL7IngresProxy = "tailscale.com/experimental-forward-cluster-traffic-via-ingress"
// Annotations set by the operator on pods to trigger restarts when the
// hostname, IP, FQDN or tailscaled config changes.
// hostname, IP, FQDN or tailscaled config changes. If you add a new
// annotation here, also add it to tailscaleManagedAnnotations var to
// ensure that it does not get removed when a ProxyClass configuration
// is applied.
podAnnotationLastSetClusterIP = "tailscale.com/operator-last-set-cluster-ip"
podAnnotationLastSetHostname = "tailscale.com/operator-last-set-hostname"
podAnnotationLastSetTailnetTargetIP = "tailscale.com/operator-last-set-ts-tailnet-target-ip"
podAnnotationLastSetTailnetTargetFQDN = "tailscale.com/operator-last-set-ts-tailnet-target-fqdn"
// podAnnotationLastSetConfigFileHash is sha256 hash of the current tailscaled configuration contents.
@@ -69,13 +97,23 @@ const (
tailscaledConfigKey = "tailscaled"
)
var (
// tailscaleManagedLabels are label keys that tailscale operator sets on StatefulSets and Pods.
tailscaleManagedLabels = []string{LabelManaged, LabelParentType, LabelParentName, LabelParentNamespace, "app"}
// tailscaleManagedAnnotations are annotation keys that tailscale operator sets on StatefulSets and Pods.
tailscaleManagedAnnotations = []string{podAnnotationLastSetClusterIP, podAnnotationLastSetTailnetTargetIP, podAnnotationLastSetTailnetTargetFQDN, podAnnotationLastSetConfigFileHash}
)
type tailscaleSTSConfig struct {
ParentResourceName string
ParentResourceUID string
ChildResourceLabels map[string]string
ServeConfig *ipn.ServeConfig
ClusterTargetIP string // ingress target
ServeConfig *ipn.ServeConfig // if serve config is set, this is a proxy for Ingress
ClusterTargetIP string // ingress target
// If set to true, operator should configure containerboot to forward
// cluster traffic via the proxy set up for Kubernetes Ingress.
ForwardClusterTrafficViaL7IngressProxy bool
TailnetTargetIP string // egress target IP
@@ -87,6 +125,8 @@ type tailscaleSTSConfig struct {
// Connector specifies a configuration of a Connector instance if that's
// what this StatefulSet should be created for.
Connector *connector
ProxyClass string
}
type connector struct {
@@ -95,10 +135,13 @@ type connector struct {
// isExitNode defines whether this Connector should act as an exit node.
isExitNode bool
}
type tsnetServer interface {
CertDomains() []string
}
type tailscaleSTSReconciler struct {
client.Client
tsnetServer *tsnet.Server
tsnetServer tsnetServer
tsClient tsClient
defaultTags []string
operatorNamespace string
@@ -214,18 +257,17 @@ const maxStatefulSetNameLength = 63 - 10 - 1
// generation will NOT result in a StatefulSet name longer than 52 chars.
// This is done because of https://github.com/kubernetes/kubernetes/issues/64023.
func statefulSetNameBase(parent string) string {
base := fmt.Sprintf("ts-%s-", parent)
// Calculate what length name GenerateName returns for this base.
generator := names.SimpleNameGenerator
generatedName := generator.GenerateName(base)
if excess := len(generatedName) - maxStatefulSetNameLength; excess > 0 {
base = base[:len(base)-excess-1] // take extra char off to make space for hyphen
base = base + "-" // re-instate hyphen
for {
generatedName := generator.GenerateName(base)
excess := len(generatedName) - maxStatefulSetNameLength
if excess <= 0 {
return base
}
base = base[:len(base)-1-excess] // cut off the excess chars
base = base + "-" // re-instate the dash
}
return base
}
func (a *tailscaleSTSReconciler) reconcileHeadlessService(ctx context.Context, logger *zap.SugaredLogger, sts *tailscaleSTSConfig) (*corev1.Service, error) {
@@ -270,9 +312,9 @@ func (a *tailscaleSTSReconciler) createOrGetSecret(ctx context.Context, logger *
authKey, hash string
)
if orig == nil {
// Secret doesn't exist yet, create one. Initially it contains
// only the Tailscale authkey, but once Tailscale starts it'll
// also store the daemon state.
// Initially it contains only tailscaled config, but when the
// proxy starts, it will also store there the state, certs and
// ACME account key.
sts, err := getSingleObject[appsv1.StatefulSet](ctx, a.Client, a.operatorNamespace, stsC.ChildResourceLabels)
if err != nil {
return "", "", err
@@ -295,17 +337,13 @@ func (a *tailscaleSTSReconciler) createOrGetSecret(ctx context.Context, logger *
return "", "", err
}
}
if !shouldDoTailscaledDeclarativeConfig(stsC) && authKey != "" {
mak.Set(&secret.StringData, "authkey", authKey)
}
if shouldDoTailscaledDeclarativeConfig(stsC) {
confFileBytes, h, err := tailscaledConfig(stsC, authKey, orig)
if err != nil {
return "", "", fmt.Errorf("error creating tailscaled config: %w", err)
}
hash = h
mak.Set(&secret.StringData, tailscaledConfigKey, string(confFileBytes))
confFileBytes, h, err := tailscaledConfig(stsC, authKey, orig)
if err != nil {
return "", "", fmt.Errorf("error creating tailscaled config: %w", err)
}
hash = h
mak.Set(&secret.StringData, tailscaledConfigKey, string(confFileBytes))
if stsC.ServeConfig != nil {
j, err := json.Marshal(stsC.ServeConfig)
if err != nil {
@@ -315,12 +353,12 @@ func (a *tailscaleSTSReconciler) createOrGetSecret(ctx context.Context, logger *
}
if orig != nil {
logger.Debugf("patching existing state Secret with values %s", secret.Data[tailscaledConfigKey])
logger.Debugf("patching the existing proxy Secret with tailscaled config %s", sanitizeConfigBytes(secret.Data[tailscaledConfigKey]))
if err := a.Patch(ctx, secret, client.MergeFrom(orig)); err != nil {
return "", "", err
}
} else {
logger.Debugf("creating new state Secret with authkey %s", secret.Data[tailscaledConfigKey])
logger.Debugf("creating a new Secret for the proxy with tailscaled config %s", sanitizeConfigBytes([]byte(secret.StringData[tailscaledConfigKey])))
if err := a.Create(ctx, secret); err != nil {
return "", "", err
}
@@ -328,6 +366,23 @@ func (a *tailscaleSTSReconciler) createOrGetSecret(ctx context.Context, logger *
return secret.Name, hash, nil
}
// sanitizeConfigBytes returns ipn.ConfigVAlpha in string form with redacted
// auth key.
func sanitizeConfigBytes(bs []byte) string {
c := &ipn.ConfigVAlpha{}
if err := json.Unmarshal(bs, c); err != nil {
return "invalid config"
}
if c.AuthKey != nil {
c.AuthKey = ptr.To("**redacted**")
}
sanitizedBytes, err := json.Marshal(c)
if err != nil {
return "invalid config"
}
return string(sanitizedBytes)
}
// DeviceInfo returns the device ID and hostname for the Tailscale device
// associated with the given labels.
func (a *tailscaleSTSReconciler) DeviceInfo(ctx context.Context, childLabels map[string]string) (id tailcfg.StableNodeID, hostname string, ips []string, err error) {
@@ -381,10 +436,10 @@ var proxyYaml []byte
var userspaceProxyYaml []byte
func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.SugaredLogger, sts *tailscaleSTSConfig, headlessSvc *corev1.Service, proxySecret, tsConfigHash string) (*appsv1.StatefulSet, error) {
var ss appsv1.StatefulSet
if sts.ServeConfig != nil {
ss := new(appsv1.StatefulSet)
if sts.ServeConfig != nil && sts.ForwardClusterTrafficViaL7IngressProxy != true { // If forwarding cluster traffic via is required we need non-userspace + NET_ADMIN + forwarding
if err := yaml.Unmarshal(userspaceProxyYaml, &ss); err != nil {
return nil, fmt.Errorf("failed to unmarshal proxy spec: %w", err)
return nil, fmt.Errorf("failed to unmarshal userspace proxy spec: %v", err)
}
} else {
if err := yaml.Unmarshal(proxyYaml, &ss); err != nil {
@@ -398,12 +453,25 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
}
}
}
container := &ss.Spec.Template.Spec.Containers[0]
pod := &ss.Spec.Template
container := &pod.Spec.Containers[0]
proxyClass := new(tsapi.ProxyClass)
if sts.ProxyClass != "" {
if err := a.Get(ctx, types.NamespacedName{Name: sts.ProxyClass}, proxyClass); err != nil {
return nil, fmt.Errorf("failed to get ProxyClass: %w", err)
}
if !tsoperator.ProxyClassIsReady(proxyClass) {
logger.Infof("ProxyClass %s specified for the proxy, but it is not (yet) in a ready state, waiting..")
return nil, nil
}
}
container.Image = a.proxyImage
ss.ObjectMeta = metav1.ObjectMeta{
Name: headlessSvc.Name,
Namespace: a.operatorNamespace,
Labels: sts.ChildResourceLabels,
}
for key, val := range sts.ChildResourceLabels {
mak.Set(&ss.ObjectMeta.Labels, key, val)
}
ss.Spec.ServiceName = headlessSvc.Name
ss.Spec.Selector = &metav1.LabelSelector{
@@ -411,7 +479,10 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
"app": sts.ParentResourceUID,
},
}
mak.Set(&ss.Spec.Template.Labels, "app", sts.ParentResourceUID)
mak.Set(&pod.Labels, "app", sts.ParentResourceUID)
for key, val := range sts.ChildResourceLabels {
pod.Labels[key] = val // sync StatefulSet labels to Pod to make it easier for users to select the Pod
}
// Generic containerboot configuration options.
container.Env = append(container.Env,
@@ -419,43 +490,36 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
Name: "TS_KUBE_SECRET",
Value: proxySecret,
},
)
if !shouldDoTailscaledDeclarativeConfig(sts) {
container.Env = append(container.Env, corev1.EnvVar{
Name: "TS_HOSTNAME",
Value: sts.Hostname,
})
// containerboot currently doesn't have a way to re-read the hostname/ip as
// it is passed via an environment variable. So we need to restart the
// container when the value changes. We do this by adding an annotation to
// the pod template that contains the last value we set.
mak.Set(&ss.Spec.Template.Annotations, podAnnotationLastSetHostname, sts.Hostname)
}
// Configure containeboot to run tailscaled with a configfile read from the state Secret.
if shouldDoTailscaledDeclarativeConfig(sts) {
mak.Set(&ss.Spec.Template.Annotations, podAnnotationLastSetConfigFileHash, tsConfigHash)
ss.Spec.Template.Spec.Volumes = append(ss.Spec.Template.Spec.Volumes, corev1.Volume{
Name: "tailscaledconfig",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{
SecretName: proxySecret,
Items: []corev1.KeyToPath{{
Key: tailscaledConfigKey,
Path: tailscaledConfigKey,
}},
},
},
})
container.VolumeMounts = append(container.VolumeMounts, corev1.VolumeMount{
Name: "tailscaledconfig",
ReadOnly: true,
MountPath: "/etc/tsconfig",
})
container.Env = append(container.Env, corev1.EnvVar{
corev1.EnvVar{
Name: "EXPERIMENTAL_TS_CONFIGFILE_PATH",
Value: "/etc/tsconfig/tailscaled",
},
)
if sts.ForwardClusterTrafficViaL7IngressProxy {
container.Env = append(container.Env, corev1.EnvVar{
Name: "EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS",
Value: "true",
})
}
// Configure containeboot to run tailscaled with a configfile read from the state Secret.
mak.Set(&ss.Spec.Template.Annotations, podAnnotationLastSetConfigFileHash, tsConfigHash)
pod.Spec.Volumes = append(ss.Spec.Template.Spec.Volumes, corev1.Volume{
Name: "tailscaledconfig",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{
SecretName: proxySecret,
Items: []corev1.KeyToPath{{
Key: tailscaledConfigKey,
Path: tailscaledConfigKey,
}},
},
},
})
container.VolumeMounts = append(container.VolumeMounts, corev1.VolumeMount{
Name: "tailscaledconfig",
ReadOnly: true,
MountPath: "/etc/tsconfig",
})
if a.tsFirewallMode != "" {
container.Env = append(container.Env, corev1.EnvVar{
@@ -463,7 +527,7 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
Value: a.tsFirewallMode,
})
}
ss.Spec.Template.Spec.PriorityClassName = a.proxyPriorityClassName
pod.Spec.PriorityClassName = a.proxyPriorityClassName
// Ingress/egress proxy configuration options.
if sts.ClusterTargetIP != "" {
@@ -494,7 +558,7 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
ReadOnly: true,
MountPath: "/etc/tailscaled",
})
ss.Spec.Template.Spec.Volumes = append(ss.Spec.Template.Spec.Volumes, corev1.Volume{
pod.Spec.Volumes = append(ss.Spec.Template.Spec.Volumes, corev1.Volume{
Name: "serve-config",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{
@@ -508,7 +572,95 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
})
}
logger.Debugf("reconciling statefulset %s/%s", ss.GetNamespace(), ss.GetName())
return createOrUpdate(ctx, a.Client, a.operatorNamespace, &ss, func(s *appsv1.StatefulSet) { s.Spec = ss.Spec })
if sts.ProxyClass != "" {
logger.Debugf("configuring proxy resources with ProxyClass %s", sts.ProxyClass)
ss = applyProxyClassToStatefulSet(proxyClass, ss)
}
updateSS := func(s *appsv1.StatefulSet) {
s.Spec = ss.Spec
s.ObjectMeta.Labels = ss.Labels
s.ObjectMeta.Annotations = ss.Annotations
}
return createOrUpdate(ctx, a.Client, a.operatorNamespace, ss, updateSS)
}
// mergeStatefulSetLabelsOrAnnots returns a map that contains all keys/values
// present in 'custom' map as well as those keys/values from the current map
// whose keys are present in the 'managed' map. The reason why this merge is
// necessary is to ensure that labels/annotations applied from a ProxyClass get removed
// if they are removed from a ProxyClass or if the ProxyClass no longer applies
// to this StatefulSet whilst any tailscale managed labels/annotations remain present.
func mergeStatefulSetLabelsOrAnnots(current, custom map[string]string, managed []string) map[string]string {
if custom == nil {
custom = make(map[string]string)
}
if current == nil {
return custom
}
for key, val := range current {
if slices.Contains(managed, key) {
custom[key] = val
}
}
return custom
}
func applyProxyClassToStatefulSet(pc *tsapi.ProxyClass, ss *appsv1.StatefulSet) *appsv1.StatefulSet {
if pc == nil || ss == nil || pc.Spec.StatefulSet == nil {
return ss
}
// Update StatefulSet metadata.
if wantsSSLabels := pc.Spec.StatefulSet.Labels; len(wantsSSLabels) > 0 {
ss.ObjectMeta.Labels = mergeStatefulSetLabelsOrAnnots(ss.ObjectMeta.Labels, wantsSSLabels, tailscaleManagedLabels)
}
if wantsSSAnnots := pc.Spec.StatefulSet.Annotations; len(wantsSSAnnots) > 0 {
ss.ObjectMeta.Annotations = mergeStatefulSetLabelsOrAnnots(ss.ObjectMeta.Annotations, wantsSSAnnots, tailscaleManagedAnnotations)
}
// Update Pod fields.
if pc.Spec.StatefulSet.Pod == nil {
return ss
}
wantsPod := pc.Spec.StatefulSet.Pod
if wantsPodLabels := wantsPod.Labels; len(wantsPodLabels) > 0 {
ss.Spec.Template.ObjectMeta.Labels = mergeStatefulSetLabelsOrAnnots(ss.Spec.Template.ObjectMeta.Labels, wantsPodLabels, tailscaleManagedLabels)
}
if wantsPodAnnots := wantsPod.Annotations; len(wantsPodAnnots) > 0 {
ss.Spec.Template.ObjectMeta.Annotations = mergeStatefulSetLabelsOrAnnots(ss.Spec.Template.ObjectMeta.Annotations, wantsPodAnnots, tailscaleManagedAnnotations)
}
ss.Spec.Template.Spec.SecurityContext = wantsPod.SecurityContext
ss.Spec.Template.Spec.ImagePullSecrets = wantsPod.ImagePullSecrets
ss.Spec.Template.Spec.NodeName = wantsPod.NodeName
ss.Spec.Template.Spec.NodeSelector = wantsPod.NodeSelector
ss.Spec.Template.Spec.Tolerations = wantsPod.Tolerations
// Update containers.
updateContainer := func(overlay *tsapi.Container, base corev1.Container) corev1.Container {
if overlay == nil {
return base
}
if overlay.SecurityContext != nil {
base.SecurityContext = overlay.SecurityContext
}
base.Resources = overlay.Resources
return base
}
for i, c := range ss.Spec.Template.Spec.Containers {
if c.Name == "tailscale" {
ss.Spec.Template.Spec.Containers[i] = updateContainer(wantsPod.TailscaleContainer, ss.Spec.Template.Spec.Containers[i])
break
}
}
if initContainers := ss.Spec.Template.Spec.InitContainers; len(initContainers) > 0 {
for i, c := range initContainers {
if c.Name == "sysctler" {
ss.Spec.Template.Spec.InitContainers[i] = updateContainer(wantsPod.TailscaleInitContainer, initContainers[i])
break
}
}
}
return ss
}
// tailscaledConfig takes a proxy config, a newly generated auth key if
@@ -516,10 +668,11 @@ func (a *tailscaleSTSReconciler) reconcileSTS(ctx context.Context, logger *zap.S
// produces returns tailscaled configuration and a hash of that configuration.
func tailscaledConfig(stsC *tailscaleSTSConfig, newAuthkey string, oldSecret *corev1.Secret) ([]byte, string, error) {
conf := ipn.ConfigVAlpha{
Version: "alpha0",
AcceptDNS: "false",
Locked: "false",
Hostname: &stsC.Hostname,
Version: "alpha0",
AcceptDNS: "false",
AcceptRoutes: "false", // AcceptRoutes defaults to true
Locked: "false",
Hostname: &stsC.Hostname,
}
if stsC.Connector != nil {
routes, err := netutil.CalcAdvertiseRoutes(stsC.Connector.routes, stsC.Connector.isExitNode)
@@ -676,10 +829,3 @@ func nameForService(svc *corev1.Service) (string, error) {
func isValidFirewallMode(m string) bool {
return m == "auto" || m == "nftables" || m == "iptables"
}
// shouldDoTailscaledDeclarativeConfig determines whether the proxy instance
// should be configured to run tailscaled only with a all config opts passed to
// tailscaled.
func shouldDoTailscaledDeclarativeConfig(stsC *tailscaleSTSConfig) bool {
return stsC.Connector != nil
}

View File

@@ -6,7 +6,20 @@
package main
import (
_ "embed"
"fmt"
"reflect"
"regexp"
"strings"
"testing"
"github.com/google/go-cmp/cmp"
appsv1 "k8s.io/api/apps/v1"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/resource"
"sigs.k8s.io/yaml"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/types/ptr"
)
// Test_statefulSetNameBase tests that parent name portion in a StatefulSet name
@@ -19,31 +32,261 @@ import (
// https://github.com/kubernetes/kubernetes/blob/v1.28.4/staging/src/k8s.io/apiserver/pkg/storage/names/generate.go#L45.
// https://github.com/kubernetes/kubernetes/pull/116430
func Test_statefulSetNameBase(t *testing.T) {
// Service name lengths can be 1 - 63 chars, be paranoid and test them all.
var b strings.Builder
for b.Len() < 63 {
if _, err := b.WriteString("a"); err != nil {
t.Fatalf("error writing to string builder: %v", err)
}
baseLength := b.Len()
if baseLength > 43 {
baseLength = 43 // currently 43 is the max base length
}
wantsNameR := regexp.MustCompile(`^ts-a{` + fmt.Sprint(baseLength) + `}-$`) // to match a string like ts-aaaa-
gotName := statefulSetNameBase(b.String())
if !wantsNameR.MatchString(gotName) {
t.Fatalf("expected string %s to match regex %s ", gotName, wantsNameR.String()) // fatal rather than error as this test is called 63 times
}
}
}
func Test_applyProxyClassToStatefulSet(t *testing.T) {
// Setup
proxyClassAllOpts := &tsapi.ProxyClass{
Spec: tsapi.ProxyClassSpec{
StatefulSet: &tsapi.StatefulSet{
Labels: map[string]string{"foo": "bar"},
Annotations: map[string]string{"foo.io/bar": "foo"},
Pod: &tsapi.Pod{
Labels: map[string]string{"bar": "foo"},
Annotations: map[string]string{"bar.io/foo": "foo"},
SecurityContext: &corev1.PodSecurityContext{
RunAsUser: ptr.To(int64(0)),
},
ImagePullSecrets: []corev1.LocalObjectReference{{Name: "docker-creds"}},
NodeName: "some-node",
NodeSelector: map[string]string{"beta.kubernetes.io/os": "linux"},
Tolerations: []corev1.Toleration{{Key: "", Operator: "Exists"}},
TailscaleContainer: &tsapi.Container{
SecurityContext: &corev1.SecurityContext{
Privileged: ptr.To(true),
},
Resources: corev1.ResourceRequirements{
Limits: corev1.ResourceList{corev1.ResourceCPU: resource.MustParse("1000m"), corev1.ResourceMemory: resource.MustParse("128Mi")},
Requests: corev1.ResourceList{corev1.ResourceCPU: resource.MustParse("500m"), corev1.ResourceMemory: resource.MustParse("64Mi")},
},
},
TailscaleInitContainer: &tsapi.Container{
SecurityContext: &corev1.SecurityContext{
Privileged: ptr.To(true),
RunAsUser: ptr.To(int64(0)),
},
Resources: corev1.ResourceRequirements{
Limits: corev1.ResourceList{corev1.ResourceCPU: resource.MustParse("1000m"), corev1.ResourceMemory: resource.MustParse("128Mi")},
Requests: corev1.ResourceList{corev1.ResourceCPU: resource.MustParse("500m"), corev1.ResourceMemory: resource.MustParse("64Mi")},
},
},
},
},
},
}
proxyClassJustLabels := &tsapi.ProxyClass{
Spec: tsapi.ProxyClassSpec{
StatefulSet: &tsapi.StatefulSet{
Labels: map[string]string{"foo": "bar"},
Annotations: map[string]string{"foo.io/bar": "foo"},
Pod: &tsapi.Pod{
Labels: map[string]string{"bar": "foo"},
Annotations: map[string]string{"bar.io/foo": "foo"},
},
},
},
}
var userspaceProxySS, nonUserspaceProxySS appsv1.StatefulSet
if err := yaml.Unmarshal(userspaceProxyYaml, &userspaceProxySS); err != nil {
t.Fatalf("unmarshaling userspace proxy template: %v", err)
}
if err := yaml.Unmarshal(proxyYaml, &nonUserspaceProxySS); err != nil {
t.Fatalf("unmarshaling non-userspace proxy template: %v", err)
}
// Set a couple additional fields so we can test that we don't
// mistakenly override those.
labels := map[string]string{
LabelManaged: "true",
LabelParentName: "foo",
}
annots := map[string]string{
podAnnotationLastSetClusterIP: "1.2.3.4",
}
env := []corev1.EnvVar{{Name: "TS_HOSTNAME", Value: "nginx"}}
userspaceProxySS.Labels = labels
userspaceProxySS.Annotations = annots
userspaceProxySS.Spec.Template.Spec.Containers[0].Env = env
nonUserspaceProxySS.ObjectMeta.Labels = labels
nonUserspaceProxySS.ObjectMeta.Annotations = annots
nonUserspaceProxySS.Spec.Template.Spec.Containers[0].Env = env
// 1. Test that a ProxyClass with all fields set gets correctly applied
// to a Statefulset built from non-userspace proxy template.
wantSS := nonUserspaceProxySS.DeepCopy()
wantSS.ObjectMeta.Labels = mergeMapKeys(wantSS.ObjectMeta.Labels, proxyClassAllOpts.Spec.StatefulSet.Labels)
wantSS.ObjectMeta.Annotations = mergeMapKeys(wantSS.ObjectMeta.Annotations, proxyClassAllOpts.Spec.StatefulSet.Annotations)
wantSS.Spec.Template.Labels = proxyClassAllOpts.Spec.StatefulSet.Pod.Labels
wantSS.Spec.Template.Annotations = proxyClassAllOpts.Spec.StatefulSet.Pod.Annotations
wantSS.Spec.Template.Spec.SecurityContext = proxyClassAllOpts.Spec.StatefulSet.Pod.SecurityContext
wantSS.Spec.Template.Spec.ImagePullSecrets = proxyClassAllOpts.Spec.StatefulSet.Pod.ImagePullSecrets
wantSS.Spec.Template.Spec.NodeName = proxyClassAllOpts.Spec.StatefulSet.Pod.NodeName
wantSS.Spec.Template.Spec.NodeSelector = proxyClassAllOpts.Spec.StatefulSet.Pod.NodeSelector
wantSS.Spec.Template.Spec.Tolerations = proxyClassAllOpts.Spec.StatefulSet.Pod.Tolerations
wantSS.Spec.Template.Spec.Containers[0].SecurityContext = proxyClassAllOpts.Spec.StatefulSet.Pod.TailscaleContainer.SecurityContext
wantSS.Spec.Template.Spec.InitContainers[0].SecurityContext = proxyClassAllOpts.Spec.StatefulSet.Pod.TailscaleInitContainer.SecurityContext
wantSS.Spec.Template.Spec.Containers[0].Resources = proxyClassAllOpts.Spec.StatefulSet.Pod.TailscaleContainer.Resources
wantSS.Spec.Template.Spec.InitContainers[0].Resources = proxyClassAllOpts.Spec.StatefulSet.Pod.TailscaleInitContainer.Resources
gotSS := applyProxyClassToStatefulSet(proxyClassAllOpts, nonUserspaceProxySS.DeepCopy())
if diff := cmp.Diff(gotSS, wantSS); diff != "" {
t.Fatalf("Unexpected result applying ProxyClass with all fields set to a StatefulSet for non-userspace proxy (-got +want):\n%s", diff)
}
// 2. Test that a ProxyClass with custom labels and annotations for
// StatefulSet and Pod set gets correctly applied to a Statefulset built
// from non-userspace proxy template.
wantSS = nonUserspaceProxySS.DeepCopy()
wantSS.ObjectMeta.Labels = mergeMapKeys(wantSS.ObjectMeta.Labels, proxyClassJustLabels.Spec.StatefulSet.Labels)
wantSS.ObjectMeta.Annotations = mergeMapKeys(wantSS.ObjectMeta.Annotations, proxyClassJustLabels.Spec.StatefulSet.Annotations)
wantSS.Spec.Template.Labels = proxyClassJustLabels.Spec.StatefulSet.Pod.Labels
wantSS.Spec.Template.Annotations = proxyClassJustLabels.Spec.StatefulSet.Pod.Annotations
gotSS = applyProxyClassToStatefulSet(proxyClassJustLabels, nonUserspaceProxySS.DeepCopy())
if diff := cmp.Diff(gotSS, wantSS); diff != "" {
t.Fatalf("Unexpected result applying ProxyClass with custom labels and annotations to a StatefulSet for non-userspace proxy (-got +want):\n%s", diff)
}
// 3. Test that a ProxyClass with all fields set gets correctly applied
// to a Statefulset built from a userspace proxy template.
wantSS = userspaceProxySS.DeepCopy()
wantSS.ObjectMeta.Labels = mergeMapKeys(wantSS.ObjectMeta.Labels, proxyClassAllOpts.Spec.StatefulSet.Labels)
wantSS.ObjectMeta.Annotations = mergeMapKeys(wantSS.ObjectMeta.Annotations, proxyClassAllOpts.Spec.StatefulSet.Annotations)
wantSS.Spec.Template.Labels = proxyClassAllOpts.Spec.StatefulSet.Pod.Labels
wantSS.Spec.Template.Annotations = proxyClassAllOpts.Spec.StatefulSet.Pod.Annotations
wantSS.Spec.Template.Spec.SecurityContext = proxyClassAllOpts.Spec.StatefulSet.Pod.SecurityContext
wantSS.Spec.Template.Spec.ImagePullSecrets = proxyClassAllOpts.Spec.StatefulSet.Pod.ImagePullSecrets
wantSS.Spec.Template.Spec.NodeName = proxyClassAllOpts.Spec.StatefulSet.Pod.NodeName
wantSS.Spec.Template.Spec.NodeSelector = proxyClassAllOpts.Spec.StatefulSet.Pod.NodeSelector
wantSS.Spec.Template.Spec.Tolerations = proxyClassAllOpts.Spec.StatefulSet.Pod.Tolerations
wantSS.Spec.Template.Spec.Containers[0].SecurityContext = proxyClassAllOpts.Spec.StatefulSet.Pod.TailscaleContainer.SecurityContext
wantSS.Spec.Template.Spec.Containers[0].Resources = proxyClassAllOpts.Spec.StatefulSet.Pod.TailscaleContainer.Resources
gotSS = applyProxyClassToStatefulSet(proxyClassAllOpts, userspaceProxySS.DeepCopy())
if diff := cmp.Diff(gotSS, wantSS); diff != "" {
t.Fatalf("Unexpected result applying ProxyClass with custom labels and annotations to a StatefulSet for a userspace proxy (-got +want):\n%s", diff)
}
// 4. Test that a ProxyClass with custom labels and annotations gets correctly applied
// to a Statefulset built from a userspace proxy template.
wantSS = userspaceProxySS.DeepCopy()
wantSS.ObjectMeta.Labels = mergeMapKeys(wantSS.ObjectMeta.Labels, proxyClassJustLabels.Spec.StatefulSet.Labels)
wantSS.ObjectMeta.Annotations = mergeMapKeys(wantSS.ObjectMeta.Annotations, proxyClassJustLabels.Spec.StatefulSet.Annotations)
wantSS.Spec.Template.Labels = proxyClassJustLabels.Spec.StatefulSet.Pod.Labels
wantSS.Spec.Template.Annotations = proxyClassJustLabels.Spec.StatefulSet.Pod.Annotations
gotSS = applyProxyClassToStatefulSet(proxyClassJustLabels, userspaceProxySS.DeepCopy())
if diff := cmp.Diff(gotSS, wantSS); diff != "" {
t.Fatalf("Unexpected result applying ProxyClass with custom labels and annotations to a StatefulSet for a userspace proxy (-got +want):\n%s", diff)
}
}
func mergeMapKeys(a, b map[string]string) map[string]string {
for key, val := range b {
a[key] = val
}
return a
}
func Test_mergeStatefulSetLabelsOrAnnots(t *testing.T) {
tests := []struct {
name string
in string
out string
name string
current map[string]string
custom map[string]string
managed []string
want map[string]string
}{
{
name: "43 chars",
in: "oidhexl9o832hcbhyg4uz6o0s7u9uae54h5k8ofs9xb",
out: "ts-oidhexl9o832hcbhyg4uz6o0s7u9uae54h5k8ofs9xb-",
name: "no custom labels specified and none present in current labels, return current labels",
current: map[string]string{LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
want: map[string]string{LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
managed: tailscaleManagedLabels,
},
{
name: "44 chars",
in: "oidhexl9o832hcbhyg4uz6o0s7u9uae54h5k8ofs9xbo",
out: "ts-oidhexl9o832hcbhyg4uz6o0s7u9uae54h5k8ofs9xb-",
name: "no custom labels specified, but some present in current labels, return tailscale managed labels only from the current labels",
current: map[string]string{"foo": "bar", "something.io/foo": "bar", LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
want: map[string]string{LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
managed: tailscaleManagedLabels,
},
{
name: "42 chars",
in: "oidhexl9o832hcbhyg4uz6o0s7u9uae54h5k8ofs9x",
out: "ts-oidhexl9o832hcbhyg4uz6o0s7u9uae54h5k8ofs9x-",
name: "custom labels specified, current labels only contain tailscale managed labels, return a union of both",
current: map[string]string{LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
custom: map[string]string{"foo": "bar", "something.io/foo": "bar"},
want: map[string]string{"foo": "bar", "something.io/foo": "bar", LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
managed: tailscaleManagedLabels,
},
{
name: "custom labels specified, current labels contain tailscale managed labels and custom labels, some of which re not present in the new custom labels, return a union of managed labels and the desired custom labels",
current: map[string]string{"foo": "bar", "bar": "baz", "app": "1234", LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
custom: map[string]string{"foo": "bar", "something.io/foo": "bar"},
want: map[string]string{"foo": "bar", "something.io/foo": "bar", "app": "1234", LabelManaged: "true", LabelParentName: "foo", LabelParentType: "svc", LabelParentNamespace: "foo"},
managed: tailscaleManagedLabels,
},
{
name: "no current labels present, return custom labels only",
custom: map[string]string{"foo": "bar", "something.io/foo": "bar"},
want: map[string]string{"foo": "bar", "something.io/foo": "bar"},
managed: tailscaleManagedLabels,
},
{
name: "no current labels present, no custom labels specified, return empty map",
want: map[string]string{},
managed: tailscaleManagedLabels,
},
{
name: "no custom annots specified and none present in current annots, return current annots",
current: map[string]string{podAnnotationLastSetClusterIP: "1.2.3.4"},
want: map[string]string{podAnnotationLastSetClusterIP: "1.2.3.4"},
managed: tailscaleManagedAnnotations,
},
{
name: "no custom annots specified, but some present in current annots, return tailscale managed annots only from the current annots",
current: map[string]string{"foo": "bar", "something.io/foo": "bar", podAnnotationLastSetClusterIP: "1.2.3.4"},
want: map[string]string{podAnnotationLastSetClusterIP: "1.2.3.4"},
managed: tailscaleManagedAnnotations,
},
{
name: "custom annots specified, current annots only contain tailscale managed annots, return a union of both",
current: map[string]string{podAnnotationLastSetClusterIP: "1.2.3.4"},
custom: map[string]string{"foo": "bar", "something.io/foo": "bar"},
want: map[string]string{"foo": "bar", "something.io/foo": "bar", podAnnotationLastSetClusterIP: "1.2.3.4"},
managed: tailscaleManagedAnnotations,
},
{
name: "custom annots specified, current annots contain tailscale managed annots and custom annots, some of which are not present in the new custom annots, return a union of managed annots and the desired custom annots",
current: map[string]string{"foo": "bar", "something.io/foo": "bar", podAnnotationLastSetClusterIP: "1.2.3.4"},
custom: map[string]string{"something.io/foo": "bar"},
want: map[string]string{"something.io/foo": "bar", podAnnotationLastSetClusterIP: "1.2.3.4"},
managed: tailscaleManagedAnnotations,
},
{
name: "no current annots present, return custom annots only",
custom: map[string]string{"foo": "bar", "something.io/foo": "bar"},
want: map[string]string{"foo": "bar", "something.io/foo": "bar"},
managed: tailscaleManagedAnnotations,
},
{
name: "no current labels present, no custom labels specified, return empty map",
want: map[string]string{},
managed: tailscaleManagedAnnotations,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
if got := statefulSetNameBase(tt.in); got != tt.out {
t.Errorf("stsNamePrefix(%s) = %q, want %s", tt.in, got, tt.out)
if got := mergeStatefulSetLabelsOrAnnots(tt.current, tt.custom, tt.managed); !reflect.DeepEqual(got, tt.want) {
t.Errorf("mergeStatefulSetLabels() = %v, want %v", got, tt.want)
}
})
}

View File

@@ -20,6 +20,8 @@ import (
"k8s.io/client-go/tools/record"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
tsoperator "tailscale.com/k8s-operator"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/util/clientmetric"
"tailscale.com/util/set"
)
@@ -155,6 +157,17 @@ func (a *ServiceReconciler) maybeProvision(ctx context.Context, logger *zap.Suga
a.logger.Error(msg)
return nil
}
proxyClass := proxyClassForObject(svc)
if proxyClass != "" {
if ready, err := proxyClassIsReady(ctx, proxyClass, a.Client); err != nil {
return fmt.Errorf("error verifying ProxyClass for Service: %w", err)
} else if !ready {
logger.Infof("ProxyClass %s specified for the Service, but is not (yet) Ready, waiting..", proxyClass)
return nil
}
}
hostname, err := nameForService(svc)
if err != nil {
return err
@@ -183,6 +196,7 @@ func (a *ServiceReconciler) maybeProvision(ctx context.Context, logger *zap.Suga
Hostname: hostname,
Tags: tags,
ChildResourceLabels: crl,
ProxyClass: proxyClass,
}
a.mu.Lock()
@@ -318,3 +332,15 @@ func (a *ServiceReconciler) tailnetTargetAnnotation(svc *corev1.Service) string
}
return svc.Annotations[annotationTailnetTargetIPOld]
}
func proxyClassForObject(o client.Object) string {
return o.GetLabels()[LabelProxyClass]
}
func proxyClassIsReady(ctx context.Context, name string, cl client.Client) (bool, error) {
proxyClass := new(tsapi.ProxyClass)
if err := cl.Get(ctx, types.NamespacedName{Name: name}, proxyClass); err != nil {
return false, fmt.Errorf("error getting ProxyClass %s: %w", name, err)
}
return tsoperator.ProxyClassIsReady(proxyClass), nil
}

View File

@@ -24,6 +24,7 @@ import (
"sigs.k8s.io/controller-runtime/pkg/reconcile"
"tailscale.com/client/tailscale"
"tailscale.com/ipn"
tsapi "tailscale.com/k8s-operator/apis/v1alpha1"
"tailscale.com/types/ptr"
"tailscale.com/util/mak"
)
@@ -31,30 +32,34 @@ import (
// confgOpts contains configuration options for creating cluster resources for
// Tailscale proxies.
type configOpts struct {
stsName string
secretName string
hostname string
namespace string
parentType string
priorityClassName string
firewallMode string
tailnetTargetIP string
tailnetTargetFQDN string
clusterTargetIP string
subnetRoutes string
isExitNode bool
shouldUseDeclarativeConfig bool // tailscaled in proxy should be configured using config file
confFileHash string
stsName string
secretName string
hostname string
namespace string
parentType string
priorityClassName string
firewallMode string
tailnetTargetIP string
tailnetTargetFQDN string
clusterTargetIP string
subnetRoutes string
isExitNode bool
confFileHash string
serveConfig *ipn.ServeConfig
shouldEnableForwardingClusterTrafficViaIngress bool
proxyClass string // configuration from the named ProxyClass should be applied to proxy resources
}
func expectedSTS(opts configOpts) *appsv1.StatefulSet {
func expectedSTS(t *testing.T, cl client.Client, opts configOpts) *appsv1.StatefulSet {
t.Helper()
tsContainer := corev1.Container{
Name: "tailscale",
Image: "tailscale/tailscale",
Env: []corev1.EnvVar{
{Name: "TS_USERSPACE", Value: "false"},
{Name: "TS_AUTH_ONCE", Value: "true"},
{Name: "POD_IP", ValueFrom: &corev1.EnvVarSource{FieldRef: &corev1.ObjectFieldSelector{APIVersion: "", FieldPath: "status.podIP"}, ResourceFieldRef: nil, ConfigMapKeyRef: nil, SecretKeyRef: nil}},
{Name: "TS_KUBE_SECRET", Value: opts.secretName},
{Name: "EXPERIMENTAL_TS_CONFIGFILE_PATH", Value: "/etc/tsconfig/tailscaled"},
},
SecurityContext: &corev1.SecurityContext{
Capabilities: &corev1.Capabilities{
@@ -63,38 +68,37 @@ func expectedSTS(opts configOpts) *appsv1.StatefulSet {
},
ImagePullPolicy: "Always",
}
if opts.shouldEnableForwardingClusterTrafficViaIngress {
tsContainer.Env = append(tsContainer.Env, corev1.EnvVar{
Name: "EXPERIMENTAL_ALLOW_PROXYING_CLUSTER_TRAFFIC_VIA_INGRESS",
Value: "true",
})
}
annots := make(map[string]string)
var volumes []corev1.Volume
if opts.shouldUseDeclarativeConfig {
volumes = []corev1.Volume{
{
Name: "tailscaledconfig",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{
SecretName: opts.secretName,
Items: []corev1.KeyToPath{
{
Key: "tailscaled",
Path: "tailscaled",
},
volumes = []corev1.Volume{
{
Name: "tailscaledconfig",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{
SecretName: opts.secretName,
Items: []corev1.KeyToPath{
{
Key: "tailscaled",
Path: "tailscaled",
},
},
},
},
}
tsContainer.VolumeMounts = []corev1.VolumeMount{{
Name: "tailscaledconfig",
ReadOnly: true,
MountPath: "/etc/tsconfig",
}}
tsContainer.Env = append(tsContainer.Env, corev1.EnvVar{
Name: "EXPERIMENTAL_TS_CONFIGFILE_PATH",
Value: "/etc/tsconfig/tailscaled",
})
},
}
tsContainer.VolumeMounts = []corev1.VolumeMount{{
Name: "tailscaledconfig",
ReadOnly: true,
MountPath: "/etc/tsconfig",
}}
if opts.confFileHash != "" {
annots["tailscale.com/operator-last-set-config-file-hash"] = opts.confFileHash
} else {
tsContainer.Env = append(tsContainer.Env, corev1.EnvVar{Name: "TS_HOSTNAME", Value: opts.hostname})
annots["tailscale.com/operator-last-set-hostname"] = opts.hostname
}
if opts.firewallMode != "" {
tsContainer.Env = append(tsContainer.Env, corev1.EnvVar{
@@ -122,7 +126,17 @@ func expectedSTS(opts configOpts) *appsv1.StatefulSet {
})
annots["tailscale.com/operator-last-set-cluster-ip"] = opts.clusterTargetIP
}
return &appsv1.StatefulSet{
if opts.serveConfig != nil {
tsContainer.Env = append(tsContainer.Env, corev1.EnvVar{
Name: "TS_SERVE_CONFIG",
Value: "/etc/tailscaled/serve-config",
})
volumes = append(volumes, corev1.Volume{
Name: "serve-config", VolumeSource: corev1.VolumeSource{Secret: &corev1.SecretVolumeSource{SecretName: opts.secretName, Items: []corev1.KeyToPath{{Path: "serve-config", Key: "serve-config"}}}},
})
tsContainer.VolumeMounts = append(tsContainer.VolumeMounts, corev1.VolumeMount{Name: "serve-config", ReadOnly: true, MountPath: "/etc/tailscaled"})
}
ss := &appsv1.StatefulSet{
TypeMeta: metav1.TypeMeta{
Kind: "StatefulSet",
APIVersion: "apps/v1",
@@ -147,7 +161,13 @@ func expectedSTS(opts configOpts) *appsv1.StatefulSet {
ObjectMeta: metav1.ObjectMeta{
Annotations: annots,
DeletionGracePeriodSeconds: ptr.To[int64](10),
Labels: map[string]string{"app": "1234-UID"},
Labels: map[string]string{
"tailscale.com/managed": "true",
"tailscale.com/parent-resource": "test",
"tailscale.com/parent-resource-ns": opts.namespace,
"tailscale.com/parent-resource-type": opts.parentType,
"app": "1234-UID",
},
},
Spec: corev1.PodSpec{
ServiceAccountName: "proxies",
@@ -169,9 +189,116 @@ func expectedSTS(opts configOpts) *appsv1.StatefulSet {
},
},
}
// If opts.proxyClass is set, retrieve the ProxyClass and apply
// configuration from that to the StatefulSet.
if opts.proxyClass != "" {
t.Logf("applying configuration from ProxyClass %s", opts.proxyClass)
proxyClass := new(tsapi.ProxyClass)
if err := cl.Get(context.Background(), types.NamespacedName{Name: opts.proxyClass}, proxyClass); err != nil {
t.Fatalf("error getting ProxyClass: %v", err)
}
return applyProxyClassToStatefulSet(proxyClass, ss)
}
return ss
}
func expectedHeadlessService(name string) *corev1.Service {
func expectedSTSUserspace(t *testing.T, cl client.Client, opts configOpts) *appsv1.StatefulSet {
t.Helper()
tsContainer := corev1.Container{
Name: "tailscale",
Image: "tailscale/tailscale",
Env: []corev1.EnvVar{
{Name: "TS_USERSPACE", Value: "true"},
{Name: "TS_KUBE_SECRET", Value: opts.secretName},
{Name: "EXPERIMENTAL_TS_CONFIGFILE_PATH", Value: "/etc/tsconfig/tailscaled"},
{Name: "TS_SERVE_CONFIG", Value: "/etc/tailscaled/serve-config"},
},
ImagePullPolicy: "Always",
VolumeMounts: []corev1.VolumeMount{
{Name: "tailscaledconfig", ReadOnly: true, MountPath: "/etc/tsconfig"},
{Name: "serve-config", ReadOnly: true, MountPath: "/etc/tailscaled"},
},
}
volumes := []corev1.Volume{
{
Name: "tailscaledconfig",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{
SecretName: opts.secretName,
Items: []corev1.KeyToPath{
{
Key: "tailscaled",
Path: "tailscaled",
},
},
},
},
},
{Name: "serve-config",
VolumeSource: corev1.VolumeSource{
Secret: &corev1.SecretVolumeSource{SecretName: opts.secretName,
Items: []corev1.KeyToPath{{Key: "serve-config", Path: "serve-config"}}}},
},
}
ss := &appsv1.StatefulSet{
TypeMeta: metav1.TypeMeta{
Kind: "StatefulSet",
APIVersion: "apps/v1",
},
ObjectMeta: metav1.ObjectMeta{
Name: opts.stsName,
Namespace: "operator-ns",
Labels: map[string]string{
"tailscale.com/managed": "true",
"tailscale.com/parent-resource": "test",
"tailscale.com/parent-resource-ns": opts.namespace,
"tailscale.com/parent-resource-type": opts.parentType,
},
},
Spec: appsv1.StatefulSetSpec{
Replicas: ptr.To[int32](1),
Selector: &metav1.LabelSelector{
MatchLabels: map[string]string{"app": "1234-UID"},
},
ServiceName: opts.stsName,
Template: corev1.PodTemplateSpec{
ObjectMeta: metav1.ObjectMeta{
DeletionGracePeriodSeconds: ptr.To[int64](10),
Labels: map[string]string{
"tailscale.com/managed": "true",
"tailscale.com/parent-resource": "test",
"tailscale.com/parent-resource-ns": opts.namespace,
"tailscale.com/parent-resource-type": opts.parentType,
"app": "1234-UID",
},
},
Spec: corev1.PodSpec{
ServiceAccountName: "proxies",
PriorityClassName: opts.priorityClassName,
Containers: []corev1.Container{tsContainer},
Volumes: volumes,
},
},
},
}
ss.Spec.Template.Annotations = map[string]string{}
if opts.confFileHash != "" {
ss.Spec.Template.Annotations["tailscale.com/operator-last-set-config-file-hash"] = opts.confFileHash
}
// If opts.proxyClass is set, retrieve the ProxyClass and apply
// configuration from that to the StatefulSet.
if opts.proxyClass != "" {
t.Logf("applying configuration from ProxyClass %s", opts.proxyClass)
proxyClass := new(tsapi.ProxyClass)
if err := cl.Get(context.Background(), types.NamespacedName{Name: opts.proxyClass}, proxyClass); err != nil {
t.Fatalf("error getting ProxyClass: %v", err)
}
return applyProxyClassToStatefulSet(proxyClass, ss)
}
return ss
}
func expectedHeadlessService(name string, parentType string) *corev1.Service {
return &corev1.Service{
TypeMeta: metav1.TypeMeta{
Kind: "Service",
@@ -185,7 +312,7 @@ func expectedHeadlessService(name string) *corev1.Service {
"tailscale.com/managed": "true",
"tailscale.com/parent-resource": "test",
"tailscale.com/parent-resource-ns": "default",
"tailscale.com/parent-resource-type": "svc",
"tailscale.com/parent-resource-type": parentType,
},
},
Spec: corev1.ServiceSpec{
@@ -199,11 +326,6 @@ func expectedHeadlessService(name string) *corev1.Service {
func expectedSecret(t *testing.T, opts configOpts) *corev1.Secret {
t.Helper()
labels := map[string]string{
"tailscale.com/managed": "true",
"tailscale.com/parent-resource": "test",
"tailscale.com/parent-resource-type": opts.parentType,
}
s := &corev1.Secret{
TypeMeta: metav1.TypeMeta{
Kind: "Secret",
@@ -214,37 +336,48 @@ func expectedSecret(t *testing.T, opts configOpts) *corev1.Secret {
Namespace: "operator-ns",
},
}
if !opts.shouldUseDeclarativeConfig {
mak.Set(&s.StringData, "authkey", "secret-authkey")
labels["tailscale.com/parent-resource-ns"] = opts.namespace
} else {
conf := &ipn.ConfigVAlpha{
Version: "alpha0",
AcceptDNS: "false",
Hostname: &opts.hostname,
Locked: "false",
AuthKey: ptr.To("secret-authkey"),
}
var routes []netip.Prefix
if opts.subnetRoutes != "" || opts.isExitNode {
r := opts.subnetRoutes
if opts.isExitNode {
r = "0.0.0.0/0,::/0," + r
}
for _, rr := range strings.Split(r, ",") {
prefix, err := netip.ParsePrefix(rr)
if err != nil {
t.Fatal(err)
}
routes = append(routes, prefix)
}
}
conf.AdvertiseRoutes = routes
b, err := json.Marshal(conf)
if opts.serveConfig != nil {
serveConfigBs, err := json.Marshal(opts.serveConfig)
if err != nil {
t.Fatalf("error marshalling tailscaled config")
t.Fatalf("error marshalling serve config: %v", err)
}
mak.Set(&s.StringData, "tailscaled", string(b))
mak.Set(&s.StringData, "serve-config", string(serveConfigBs))
}
conf := &ipn.ConfigVAlpha{
Version: "alpha0",
AcceptDNS: "false",
Hostname: &opts.hostname,
Locked: "false",
AuthKey: ptr.To("secret-authkey"),
AcceptRoutes: "false",
}
var routes []netip.Prefix
if opts.subnetRoutes != "" || opts.isExitNode {
r := opts.subnetRoutes
if opts.isExitNode {
r = "0.0.0.0/0,::/0," + r
}
for _, rr := range strings.Split(r, ",") {
prefix, err := netip.ParsePrefix(rr)
if err != nil {
t.Fatal(err)
}
routes = append(routes, prefix)
}
}
conf.AdvertiseRoutes = routes
b, err := json.Marshal(conf)
if err != nil {
t.Fatalf("error marshalling tailscaled config")
}
mak.Set(&s.StringData, "tailscaled", string(b))
labels := map[string]string{
"tailscale.com/managed": "true",
"tailscale.com/parent-resource": "test",
"tailscale.com/parent-resource-ns": "default",
"tailscale.com/parent-resource-type": opts.parentType,
}
if opts.parentType == "connector" {
labels["tailscale.com/parent-resource-ns"] = "" // Connector is cluster scoped
}
s.Labels = labels
@@ -306,7 +439,13 @@ func mustUpdateStatus[T any, O ptrObject[T]](t *testing.T, client client.Client,
}
}
func expectEqual[T any, O ptrObject[T]](t *testing.T, client client.Client, want O) {
// expectEqual accepts a Kubernetes object and a Kubernetes client. It tests
// whether an object with equivalent contents can be retrieved by the passed
// client. If you want to NOT test some object fields for equality, ensure that
// they are not present in the passed object and use the modify func to remove
// them from the cluster object. If no such modifications are needed, you can
// pass nil in place of the modify function.
func expectEqual[T any, O ptrObject[T]](t *testing.T, client client.Client, want O, modify func(O)) {
t.Helper()
got := O(new(T))
if err := client.Get(context.Background(), types.NamespacedName{
@@ -320,6 +459,9 @@ func expectEqual[T any, O ptrObject[T]](t *testing.T, client client.Client, want
// so just remove it from both got and want.
got.SetResourceVersion("")
want.SetResourceVersion("")
if modify != nil {
modify(got)
}
if diff := cmp.Diff(got, want); diff != "" {
t.Fatalf("unexpected object (-got +want):\n%s", diff)
}
@@ -378,6 +520,13 @@ type fakeTSClient struct {
keyRequests []tailscale.KeyCapabilities
deleted []string
}
type fakeTSNetServer struct {
certDomains []string
}
func (f *fakeTSNetServer) CertDomains() []string {
return f.certDomains
}
func (c *fakeTSClient) CreateKey(ctx context.Context, caps tailscale.KeyCapabilities) (string, *tailscale.Key, error) {
c.Lock()
@@ -409,3 +558,11 @@ func (c *fakeTSClient) Deleted() []string {
defer c.Unlock()
return c.deleted
}
// removeHashAnnotation can be used to remove declarative tailscaled config hash
// annotation from proxy StatefulSets to make the tests more maintainable (so
// that we don't have to change the annotation in each test case after any
// change to the configfile contents).
func removeHashAnnotation(sts *appsv1.StatefulSet) {
delete(sts.Spec.Template.Annotations, podAnnotationLastSetConfigFileHash)
}

View File

@@ -12,7 +12,7 @@ import (
"net/netip"
"slices"
"inet.af/tcpproxy"
"github.com/inetaf/tcpproxy"
"tailscale.com/net/netutil"
)

View File

@@ -2,9 +2,7 @@ tailscale.com/cmd/stund dependencies: (generated by github.com/tailscale/depawar
github.com/beorn7/perks/quantile from github.com/prometheus/client_golang/prometheus
💣 github.com/cespare/xxhash/v2 from github.com/prometheus/client_golang/prometheus
github.com/golang/protobuf/proto from github.com/matttproud/golang_protobuf_extensions/pbutil
github.com/google/uuid from tailscale.com/tsweb
github.com/matttproud/golang_protobuf_extensions/pbutil from github.com/prometheus/common/expfmt
💣 github.com/prometheus/client_golang/prometheus from tailscale.com/tsweb/promvarz
github.com/prometheus/client_golang/prometheus/internal from github.com/prometheus/client_golang/prometheus
github.com/prometheus/client_model/go from github.com/prometheus/client_golang/prometheus+
@@ -16,8 +14,9 @@ tailscale.com/cmd/stund dependencies: (generated by github.com/tailscale/depawar
LD github.com/prometheus/procfs/internal/util from github.com/prometheus/procfs
💣 go4.org/mem from tailscale.com/metrics+
go4.org/netipx from tailscale.com/net/tsaddr
google.golang.org/protobuf/encoding/prototext from github.com/golang/protobuf/proto+
google.golang.org/protobuf/encoding/protowire from github.com/golang/protobuf/proto+
google.golang.org/protobuf/encoding/protodelim from github.com/prometheus/common/expfmt
google.golang.org/protobuf/encoding/prototext from github.com/prometheus/common/expfmt+
google.golang.org/protobuf/encoding/protowire from google.golang.org/protobuf/encoding/protodelim+
google.golang.org/protobuf/internal/descfmt from google.golang.org/protobuf/internal/filedesc
google.golang.org/protobuf/internal/descopts from google.golang.org/protobuf/internal/filedesc+
google.golang.org/protobuf/internal/detrand from google.golang.org/protobuf/internal/descfmt+
@@ -36,13 +35,11 @@ tailscale.com/cmd/stund dependencies: (generated by github.com/tailscale/depawar
google.golang.org/protobuf/internal/set from google.golang.org/protobuf/encoding/prototext
💣 google.golang.org/protobuf/internal/strs from google.golang.org/protobuf/encoding/prototext+
google.golang.org/protobuf/internal/version from google.golang.org/protobuf/runtime/protoimpl
google.golang.org/protobuf/proto from github.com/golang/protobuf/proto+
google.golang.org/protobuf/reflect/protodesc from github.com/golang/protobuf/proto
💣 google.golang.org/protobuf/reflect/protoreflect from github.com/golang/protobuf/proto+
google.golang.org/protobuf/reflect/protoregistry from github.com/golang/protobuf/proto+
google.golang.org/protobuf/runtime/protoiface from github.com/golang/protobuf/proto+
google.golang.org/protobuf/runtime/protoimpl from github.com/golang/protobuf/proto+
google.golang.org/protobuf/types/descriptorpb from google.golang.org/protobuf/reflect/protodesc
google.golang.org/protobuf/proto from github.com/prometheus/client_golang/prometheus+
💣 google.golang.org/protobuf/reflect/protoreflect from github.com/prometheus/client_model/go+
google.golang.org/protobuf/reflect/protoregistry from google.golang.org/protobuf/encoding/prototext+
google.golang.org/protobuf/runtime/protoiface from google.golang.org/protobuf/internal/impl+
google.golang.org/protobuf/runtime/protoimpl from github.com/prometheus/client_model/go+
google.golang.org/protobuf/types/known/timestamppb from github.com/prometheus/client_golang/prometheus+
tailscale.com from tailscale.com/version
tailscale.com/envknob from tailscale.com/tsweb+
@@ -65,7 +62,7 @@ tailscale.com/cmd/stund dependencies: (generated by github.com/tailscale/depawar
tailscale.com/types/structs from tailscale.com/tailcfg+
tailscale.com/types/tkatype from tailscale.com/tailcfg+
tailscale.com/types/views from tailscale.com/net/tsaddr+
tailscale.com/util/cmpx from tailscale.com/tailcfg+
tailscale.com/util/ctxkey from tailscale.com/tsweb+
L 💣 tailscale.com/util/dirwalk from tailscale.com/metrics
tailscale.com/util/dnsname from tailscale.com/tailcfg
tailscale.com/util/lineread from tailscale.com/version/distro
@@ -99,9 +96,9 @@ tailscale.com/cmd/stund dependencies: (generated by github.com/tailscale/depawar
golang.org/x/text/unicode/norm from golang.org/x/net/idna
bufio from compress/flate+
bytes from bufio+
cmp from slices
cmp from slices+
compress/flate from compress/gzip
compress/gzip from github.com/golang/protobuf/proto+
compress/gzip from google.golang.org/protobuf/internal/impl+
container/list from crypto/tls+
context from crypto/tls+
crypto from crypto/ecdh+
@@ -146,7 +143,7 @@ tailscale.com/cmd/stund dependencies: (generated by github.com/tailscale/depawar
html from net/http/pprof+
io from bufio+
io/fs from crypto/x509+
io/ioutil from github.com/golang/protobuf/proto+
io/ioutil from google.golang.org/protobuf/internal/impl
log from expvar+
log/internal from log
maps from tailscale.com/tailcfg+

View File

@@ -147,6 +147,8 @@ change in the future.
switch {
case slices.Contains(args, "debug"):
rootCmd.Subcommands = append(rootCmd.Subcommands, debugCmd)
case slices.Contains(args, "share"):
rootCmd.Subcommands = append(rootCmd.Subcommands, shareCmd)
}
if runtime.GOOS == "linux" && distro.Get() == distro.Synology {
rootCmd.Subcommands = append(rootCmd.Subcommands, configureHostCmd)

View File

@@ -5,6 +5,7 @@ package cli
import (
"bytes"
stdcmp "cmp"
"encoding/json"
"flag"
"fmt"
@@ -24,7 +25,6 @@ import (
"tailscale.com/types/logger"
"tailscale.com/types/persist"
"tailscale.com/types/preftype"
"tailscale.com/util/cmpx"
"tailscale.com/version/distro"
)
@@ -758,7 +758,7 @@ func TestPrefsFromUpArgs(t *testing.T) {
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
var warnBuf tstest.MemLogger
goos := cmpx.Or(tt.goos, "linux")
goos := stdcmp.Or(tt.goos, "linux")
st := tt.st
if st == nil {
st = new(ipnstate.Status)
@@ -802,7 +802,7 @@ func TestPrefFlagMapping(t *testing.T) {
}
}
prefType := reflect.TypeOf(ipn.Prefs{})
prefType := reflect.TypeFor[ipn.Prefs]()
for i := 0; i < prefType.NumField(); i++ {
prefName := prefType.Field(i).Name
if prefHasFlag[prefName] {
@@ -829,6 +829,10 @@ func TestPrefFlagMapping(t *testing.T) {
// Handled by TS_DEBUG_FIREWALL_MODE env var, we don't want to have
// a CLI flag for this. The Pref is used by c2n.
continue
case "TailFSShares":
// Handled by the tailscale share subcommand, we don't want a CLI
// flag for this.
continue
}
t.Errorf("unexpected new ipn.Pref field %q is not handled by up.go (see addPrefFlagMapping and checkForAccidentalSettingReverts)", prefName)
}

View File

@@ -119,7 +119,7 @@ func updateKubeconfig(cfgYaml []byte, fqdn string) ([]byte, error) {
var clusters []any
if cm, ok := cfg["clusters"]; ok {
clusters = cm.([]any)
clusters, _ = cm.([]any)
}
cfg["clusters"] = appendOrSetNamed(clusters, fqdn, map[string]any{
"name": fqdn,
@@ -130,7 +130,7 @@ func updateKubeconfig(cfgYaml []byte, fqdn string) ([]byte, error) {
var users []any
if um, ok := cfg["users"]; ok {
users = um.([]any)
users, _ = um.([]any)
}
cfg["users"] = appendOrSetNamed(users, "tailscale-auth", map[string]any{
// We just need one of these, and can reuse it for all clusters.
@@ -144,7 +144,7 @@ func updateKubeconfig(cfgYaml []byte, fqdn string) ([]byte, error) {
var contexts []any
if cm, ok := cfg["contexts"]; ok {
contexts = cm.([]any)
contexts, _ = cm.([]any)
}
cfg["contexts"] = appendOrSetNamed(contexts, fqdn, map[string]any{
"name": fqdn,

View File

@@ -48,6 +48,31 @@ contexts:
current-context: foo.tail-scale.ts.net
kind: Config
users:
- name: tailscale-auth
user:
token: unused`,
},
{
name: "all configs, clusters, users have been deleted",
in: `apiVersion: v1
clusters: null
contexts: null
kind: Config
current-context: some-non-existent-cluster
users: null`,
want: `apiVersion: v1
clusters:
- cluster:
server: https://foo.tail-scale.ts.net
name: foo.tail-scale.ts.net
contexts:
- context:
cluster: foo.tail-scale.ts.net
user: tailscale-auth
name: foo.tail-scale.ts.net
current-context: foo.tail-scale.ts.net
kind: Config
users:
- name: tailscale-auth
user:
token: unused`,

View File

@@ -130,14 +130,14 @@ var debugCmd = &ffcli.Command{
ShortHelp: "force a magicsock rebind",
},
{
Name: "derp-set-homeless",
Name: "derp-set-on-demand",
Exec: localAPIAction("derp-set-homeless"),
ShortHelp: "enable DERP homeless mode (breaks reachablility)",
ShortHelp: "enable DERP on-demand mode (breaks reachability)",
},
{
Name: "derp-unset-homeless",
Name: "derp-unset-on-demand",
Exec: localAPIAction("derp-unset-homeless"),
ShortHelp: "disable DERP homeless mode",
ShortHelp: "disable DERP on-demand mode",
},
{
Name: "break-tcp-conns",

View File

@@ -23,6 +23,8 @@ import (
var exitNodeCmd = &ffcli.Command{
Name: "exit-node",
ShortUsage: "exit-node [flags]",
ShortHelp: "Show machines on your tailnet configured as exit nodes",
LongHelp: "Show machines on your tailnet configured as exit nodes",
Subcommands: []*ffcli.Command{
{
Name: "list",

View File

@@ -15,7 +15,6 @@ import (
"github.com/peterbourgon/ff/v3/ffcli"
"tailscale.com/ipn"
"tailscale.com/tailcfg"
"tailscale.com/util/mak"
)
var funnelCmd = func() *ffcli.Command {
@@ -114,15 +113,8 @@ func (e *serveEnv) runFunnel(ctx context.Context, args []string) error {
// Nothing to do.
return nil
}
if on {
mak.Set(&sc.AllowFunnel, hp, true)
} else {
delete(sc.AllowFunnel, hp)
// clear map mostly for testing
if len(sc.AllowFunnel) == 0 {
sc.AllowFunnel = nil
}
}
sc.SetFunnel(dnsName, port, on)
if err := e.lc.SetServeConfig(ctx, sc); err != nil {
return err
}

View File

@@ -27,7 +27,6 @@ import (
"tailscale.com/ipn"
"tailscale.com/ipn/ipnstate"
"tailscale.com/tailcfg"
"tailscale.com/util/mak"
"tailscale.com/version"
)
@@ -357,35 +356,12 @@ func (e *serveEnv) handleWebServe(ctx context.Context, srvPort uint16, useTLS bo
if err != nil {
return err
}
hp := ipn.HostPort(net.JoinHostPort(dnsName, strconv.Itoa(int(srvPort))))
if sc.IsTCPForwardingOnPort(srvPort) {
fmt.Fprintf(os.Stderr, "error: cannot serve web; already serving TCP\n")
return errHelp
}
mak.Set(&sc.TCP, srvPort, &ipn.TCPPortHandler{HTTPS: useTLS, HTTP: !useTLS})
if _, ok := sc.Web[hp]; !ok {
mak.Set(&sc.Web, hp, new(ipn.WebServerConfig))
}
mak.Set(&sc.Web[hp].Handlers, mount, h)
for k, v := range sc.Web[hp].Handlers {
if v == h {
continue
}
// If the new mount point ends in / and another mount point
// shares the same prefix, remove the other handler.
// (e.g. /foo/ overwrites /foo)
// The opposite example is also handled.
m1 := strings.TrimSuffix(mount, "/")
m2 := strings.TrimSuffix(k, "/")
if m1 == m2 {
delete(sc.Web[hp].Handlers, k)
continue
}
}
sc.SetWebHandler(h, dnsName, srvPort, mount, useTLS)
if !reflect.DeepEqual(cursc, sc) {
if err := e.lc.SetServeConfig(ctx, sc); err != nil {
@@ -444,19 +420,7 @@ func (e *serveEnv) handleWebServeRemove(ctx context.Context, srvPort uint16, mou
if !sc.WebHandlerExists(hp, mount) {
return errors.New("error: handler does not exist")
}
// delete existing handler, then cascade delete if empty
delete(sc.Web[hp].Handlers, mount)
if len(sc.Web[hp].Handlers) == 0 {
delete(sc.Web, hp)
delete(sc.TCP, srvPort)
}
// clear empty maps mostly for testing
if len(sc.Web) == 0 {
sc.Web = nil
}
if len(sc.TCP) == 0 {
sc.TCP = nil
}
sc.RemoveWebHandler(dnsName, srvPort, []string{mount}, false)
if err := e.lc.SetServeConfig(ctx, sc); err != nil {
return err
}
@@ -592,15 +556,12 @@ func (e *serveEnv) handleTCPServe(ctx context.Context, srcType string, srcPort u
return fmt.Errorf("cannot serve TCP; already serving web on %d", srcPort)
}
mak.Set(&sc.TCP, srcPort, &ipn.TCPPortHandler{TCPForward: fwdAddr})
dnsName, err := e.getSelfDNSName(ctx)
if err != nil {
return err
}
if terminateTLS {
sc.TCP[srcPort].TerminateTLS = dnsName
}
sc.SetTCPForwarding(srcPort, fwdAddr, terminateTLS, dnsName)
if !reflect.DeepEqual(cursc, sc) {
if err := e.lc.SetServeConfig(ctx, sc); err != nil {
@@ -626,11 +587,7 @@ func (e *serveEnv) handleTCPServeRemove(ctx context.Context, src uint16) error {
return fmt.Errorf("unable to remove; serving web, not TCP forwarding on serve port %d", src)
}
if ph := sc.GetTCPPortHandler(src); ph != nil {
delete(sc.TCP, src)
// clear map mostly for testing
if len(sc.TCP) == 0 {
sc.TCP = nil
}
sc.RemoveTCPForwarding(src)
return e.lc.SetServeConfig(ctx, sc)
}
return errors.New("error: serve config does not exist")
@@ -642,6 +599,9 @@ func (e *serveEnv) handleTCPServeRemove(ctx context.Context, src uint16) error {
// Examples:
// - tailscale status
// - tailscale status --json
//
// TODO(tyler,marwan,sonia): `status` should also report foreground configs,
// currently only reports background config.
func (e *serveEnv) runServeStatus(ctx context.Context, args []string) error {
sc, err := e.lc.GetServeConfig(ctx)
if err != nil {

View File

@@ -21,6 +21,7 @@ import (
"tailscale.com/ipn"
"tailscale.com/ipn/ipnstate"
"tailscale.com/tailcfg"
"tailscale.com/tstest"
"tailscale.com/types/logger"
)
@@ -807,9 +808,11 @@ func TestVerifyFunnelEnabled(t *testing.T) {
lc.setQueryFeatureResponse(tt.queryFeatureResponse)
if tt.caps != nil {
oldCaps := fakeStatus.Self.Capabilities
defer func() { fakeStatus.Self.Capabilities = oldCaps }() // reset after test
fakeStatus.Self.Capabilities = tt.caps
cm := make(tailcfg.NodeCapMap)
for _, c := range tt.caps {
cm[c] = nil
}
tstest.Replace(t, &fakeStatus.Self.CapMap, cm)
}
defer func() {
@@ -853,8 +856,11 @@ type fakeLocalServeClient struct {
var fakeStatus = &ipnstate.Status{
BackendState: ipn.Running.String(),
Self: &ipnstate.PeerStatus{
DNSName: "foo.test.ts.net",
Capabilities: []tailcfg.NodeCapability{tailcfg.NodeAttrFunnel, tailcfg.CapabilityFunnelPorts + "?ports=443,8443"},
DNSName: "foo.test.ts.net",
CapMap: tailcfg.NodeCapMap{
tailcfg.NodeAttrFunnel: nil,
tailcfg.CapabilityFunnelPorts + "?ports=443,8443": nil,
},
},
}

View File

@@ -18,7 +18,6 @@ import (
"os/signal"
"path"
"path/filepath"
"slices"
"sort"
"strconv"
"strings"
@@ -272,7 +271,7 @@ func (e *serveEnv) runServeCombined(subcmd serveMode) execFunc {
// if foreground mode, create a WatchIPNBus session
// and use the nested config for all following operations
// TODO(marwan-at-work): nested-config validations should happen here or previous to this point.
watcher, err = e.lc.WatchIPNBus(ctx, ipn.NotifyInitialState)
watcher, err = e.lc.WatchIPNBus(ctx, ipn.NotifyInitialState|ipn.NotifyNoPrivateKeys)
if err != nil {
return err
}
@@ -334,7 +333,7 @@ func (e *serveEnv) runServeCombined(subcmd serveMode) execFunc {
const backgroundExistsMsg = "background configuration already exists, use `tailscale %s --%s=%d off` to remove the existing configuration"
func (e *serveEnv) validateConfig(sc *ipn.ServeConfig, port uint16, wantServe serveType) error {
sc, isFg := findConfig(sc, port)
sc, isFg := sc.FindConfig(port)
if sc == nil {
return nil
}
@@ -366,24 +365,6 @@ func serveFromPortHandler(tcp *ipn.TCPPortHandler) serveType {
}
}
// findConfig finds a config that contains the given port, which can be
// the top level background config or an inner foreground one. The second
// result is true if it's foreground
func findConfig(sc *ipn.ServeConfig, port uint16) (*ipn.ServeConfig, bool) {
if sc == nil {
return nil, false
}
if _, ok := sc.TCP[port]; ok {
return sc, false
}
for _, sc := range sc.Foreground {
if _, ok := sc.TCP[port]; ok {
return sc, true
}
}
return nil, false
}
func (e *serveEnv) setServe(sc *ipn.ServeConfig, st *ipnstate.Status, dnsName string, srvType serveType, srvPort uint16, mount string, target string, allowFunnel bool) error {
// update serve config based on the type
switch srvType {
@@ -516,9 +497,9 @@ func (e *serveEnv) applyWebServe(sc *ipn.ServeConfig, dnsName string, srvPort ui
}
h.Text = text
case filepath.IsAbs(target):
if version.IsSandboxedMacOS() {
// don't allow path serving for now on macOS (2022-11-15)
return errors.New("path serving is not supported if sandboxed on macOS")
if version.IsMacAppStore() || version.IsMacSys() {
// The Tailscale network extension cannot serve arbitrary paths on macOS due to sandbox restrictions (2024-03-26)
return errors.New("Path serving is not supported on macOS due to sandbox restrictions. To use Tailscale Serve on macOS, switch to the open-source tailscaled distribution. See https://tailscale.com/kb/1065/macos-variants for more information.")
}
target = filepath.Clean(target)
@@ -535,7 +516,7 @@ func (e *serveEnv) applyWebServe(sc *ipn.ServeConfig, dnsName string, srvPort ui
}
h.Path = target
default:
t, err := expandProxyTargetDev(target, []string{"http", "https", "https+insecure"}, "http")
t, err := ipn.ExpandProxyTargetValue(target, []string{"http", "https", "https+insecure"}, "http")
if err != nil {
return err
}
@@ -547,29 +528,7 @@ func (e *serveEnv) applyWebServe(sc *ipn.ServeConfig, dnsName string, srvPort ui
return errors.New("cannot serve web; already serving TCP")
}
mak.Set(&sc.TCP, srvPort, &ipn.TCPPortHandler{HTTPS: useTLS, HTTP: !useTLS})
hp := ipn.HostPort(net.JoinHostPort(dnsName, strconv.Itoa(int(srvPort))))
if _, ok := sc.Web[hp]; !ok {
mak.Set(&sc.Web, hp, new(ipn.WebServerConfig))
}
mak.Set(&sc.Web[hp].Handlers, mount, h)
// TODO: handle multiple web handlers from foreground mode
for k, v := range sc.Web[hp].Handlers {
if v == h {
continue
}
// If the new mount point ends in / and another mount point
// shares the same prefix, remove the other handler.
// (e.g. /foo/ overwrites /foo)
// The opposite example is also handled.
m1 := strings.TrimSuffix(mount, "/")
m2 := strings.TrimSuffix(k, "/")
if m1 == m2 {
delete(sc.Web[hp].Handlers, k)
}
}
sc.SetWebHandler(h, dnsName, srvPort, mount, useTLS)
return nil
}
@@ -585,7 +544,7 @@ func (e *serveEnv) applyTCPServe(sc *ipn.ServeConfig, dnsName string, srcType se
return fmt.Errorf("invalid TCP target %q", target)
}
targetURL, err := expandProxyTargetDev(target, []string{"tcp"}, "tcp")
targetURL, err := ipn.ExpandProxyTargetValue(target, []string{"tcp"}, "tcp")
if err != nil {
return fmt.Errorf("unable to expand target: %v", err)
}
@@ -600,11 +559,7 @@ func (e *serveEnv) applyTCPServe(sc *ipn.ServeConfig, dnsName string, srcType se
return fmt.Errorf("cannot serve TCP; already serving web on %d", srcPort)
}
mak.Set(&sc.TCP, srcPort, &ipn.TCPPortHandler{TCPForward: dstURL.Host})
if terminateTLS {
sc.TCP[srcPort].TerminateTLS = dnsName
}
sc.SetTCPForwarding(srcPort, dstURL.Host, terminateTLS, dnsName)
return nil
}
@@ -618,14 +573,10 @@ func (e *serveEnv) applyFunnel(sc *ipn.ServeConfig, dnsName string, srvPort uint
sc = new(ipn.ServeConfig)
}
// TODO: should ensure there is no other conflicting funnel
// TODO: add error handling for if toggling for existing sc
if allowFunnel {
mak.Set(&sc.AllowFunnel, hp, true)
} else if _, exists := sc.AllowFunnel[hp]; exists {
fmt.Fprintf(e.stderr(), "Removing Funnel for %s\n", hp)
delete(sc.AllowFunnel, hp)
if _, exists := sc.AllowFunnel[hp]; exists && !allowFunnel {
fmt.Fprintf(e.stderr(), "Removing Funnel for %s:%s\n", dnsName, hp)
}
sc.SetFunnel(dnsName, srvPort, allowFunnel)
}
// unsetServe removes the serve config for the given serve port.
@@ -814,34 +765,7 @@ func (e *serveEnv) removeWebServe(sc *ipn.ServeConfig, dnsName string, srvPort u
}
}
// delete existing handler, then cascade delete if empty
for _, m := range mounts {
delete(sc.Web[hp].Handlers, m)
}
if len(sc.Web[hp].Handlers) == 0 {
delete(sc.Web, hp)
delete(sc.AllowFunnel, hp)
delete(sc.TCP, srvPort)
}
// clear empty maps mostly for testing
if len(sc.Web) == 0 {
sc.Web = nil
}
if len(sc.TCP) == 0 {
sc.TCP = nil
}
// disable funnel if no remaining mounts exist for the serve port
if sc.Web == nil && sc.TCP == nil {
delete(sc.AllowFunnel, hp)
}
if len(sc.AllowFunnel) == 0 {
sc.AllowFunnel = nil
}
sc.RemoveWebHandler(dnsName, srvPort, mounts, true)
return nil
}
@@ -857,68 +781,10 @@ func (e *serveEnv) removeTCPServe(sc *ipn.ServeConfig, src uint16) error {
if sc.IsServingWeb(src) {
return fmt.Errorf("unable to remove; serving web, not TCP forwarding on serve port %d", src)
}
delete(sc.TCP, src)
// clear map mostly for testing
if len(sc.TCP) == 0 {
sc.TCP = nil
}
sc.RemoveTCPForwarding(src)
return nil
}
// expandProxyTargetDev expands the supported target values to be proxied
// allowing for input values to be a port number, a partial URL, or a full URL
// including a path.
//
// examples:
// - 3000
// - localhost:3000
// - tcp://localhost:3000
// - http://localhost:3000
// - https://localhost:3000
// - https-insecure://localhost:3000
// - https-insecure://localhost:3000/foo
func expandProxyTargetDev(target string, supportedSchemes []string, defaultScheme string) (string, error) {
const host = "127.0.0.1"
// support target being a port number
if port, err := strconv.ParseUint(target, 10, 16); err == nil {
return fmt.Sprintf("%s://%s:%d", defaultScheme, host, port), nil
}
// prepend scheme if not present
if !strings.Contains(target, "://") {
target = defaultScheme + "://" + target
}
// make sure we can parse the target
u, err := url.ParseRequestURI(target)
if err != nil {
return "", fmt.Errorf("invalid URL %w", err)
}
// ensure a supported scheme
if !slices.Contains(supportedSchemes, u.Scheme) {
return "", fmt.Errorf("must be a URL starting with one of the supported schemes: %v", supportedSchemes)
}
// validate the host.
switch u.Hostname() {
case "localhost", "127.0.0.1":
default:
return "", errors.New("only localhost or 127.0.0.1 proxies are currently supported")
}
// validate the port
port, err := strconv.ParseUint(u.Port(), 10, 16)
if err != nil || port == 0 {
return "", fmt.Errorf("invalid port %q", u.Port())
}
u.Host = fmt.Sprintf("%s:%d", host, port)
return u.String(), nil
}
// cleanURLPath ensures the path is clean and has a leading "/".
func cleanURLPath(urlPath string) (string, error) {
if urlPath == "" {

View File

@@ -1041,63 +1041,6 @@ func TestSrcTypeFromFlags(t *testing.T) {
}
}
func TestExpandProxyTargetDev(t *testing.T) {
tests := []struct {
name string
input string
defaultScheme string
supportedSchemes []string
expected string
wantErr bool
}{
{name: "port-only", input: "8080", expected: "http://127.0.0.1:8080"},
{name: "hostname+port", input: "localhost:8080", expected: "http://127.0.0.1:8080"},
{name: "convert-localhost", input: "http://localhost:8080", expected: "http://127.0.0.1:8080"},
{name: "no-change", input: "http://127.0.0.1:8080", expected: "http://127.0.0.1:8080"},
{name: "include-path", input: "http://127.0.0.1:8080/foo", expected: "http://127.0.0.1:8080/foo"},
{name: "https-scheme", input: "https://localhost:8080", expected: "https://127.0.0.1:8080"},
{name: "https+insecure-scheme", input: "https+insecure://localhost:8080", expected: "https+insecure://127.0.0.1:8080"},
{name: "change-default-scheme", input: "localhost:8080", defaultScheme: "https", expected: "https://127.0.0.1:8080"},
{name: "change-supported-schemes", input: "localhost:8080", defaultScheme: "tcp", supportedSchemes: []string{"tcp"}, expected: "tcp://127.0.0.1:8080"},
// errors
{name: "invalid-port", input: "localhost:9999999", wantErr: true},
{name: "unsupported-scheme", input: "ftp://localhost:8080", expected: "", wantErr: true},
{name: "not-localhost", input: "https://tailscale.com:8080", expected: "", wantErr: true},
{name: "empty-input", input: "", expected: "", wantErr: true},
}
for _, tt := range tests {
defaultScheme := "http"
supportedSchemes := []string{"http", "https", "https+insecure"}
if tt.supportedSchemes != nil {
supportedSchemes = tt.supportedSchemes
}
if tt.defaultScheme != "" {
defaultScheme = tt.defaultScheme
}
t.Run(tt.name, func(t *testing.T) {
actual, err := expandProxyTargetDev(tt.input, supportedSchemes, defaultScheme)
if tt.wantErr == true && err == nil {
t.Errorf("Expected an error but got none")
return
}
if tt.wantErr == false && err != nil {
t.Errorf("Got an error, but didn't expect one: %v", err)
return
}
if actual != tt.expected {
t.Errorf("Got: %q; expected: %q", actual, tt.expected)
}
})
}
}
func TestCleanURLPath(t *testing.T) {
tests := []struct {
input string

View File

@@ -75,7 +75,7 @@ func newSetFlagSet(goos string, setArgs *setArgsT) *flag.FlagSet {
setf.BoolVar(&setArgs.updateCheck, "update-check", true, "notify about available Tailscale updates")
setf.BoolVar(&setArgs.updateApply, "auto-update", false, "automatically update to the latest available version")
setf.BoolVar(&setArgs.postureChecking, "posture-checking", false, "HIDDEN: allow management plane to gather device posture information")
setf.BoolVar(&setArgs.runWebClient, "webclient", false, "run a web interface for managing this node, served over Tailscale at port 5252")
setf.BoolVar(&setArgs.runWebClient, "webclient", false, "expose the web interface for managing this node over Tailscale at port 5252")
if safesocket.GOOSUsesPeerCreds(goos) {
setf.StringVar(&setArgs.opUser, "operator", "", "Unix username to allow to operate on tailscaled without sudo")

243
cmd/tailscale/cli/share.go Normal file
View File

@@ -0,0 +1,243 @@
// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
package cli
import (
"context"
"errors"
"fmt"
"strings"
"github.com/peterbourgon/ff/v3/ffcli"
"tailscale.com/tailfs"
)
const (
shareSetUsage = "share set <name> <path>"
shareRenameUsage = "share rename <oldname> <newname>"
shareRemoveUsage = "share remove <name>"
shareListUsage = "share list"
)
var shareCmd = &ffcli.Command{
Name: "share",
ShortHelp: "Share a directory with your tailnet",
ShortUsage: strings.Join([]string{
shareSetUsage,
shareRemoveUsage,
shareListUsage,
}, "\n "),
LongHelp: buildShareLongHelp(),
UsageFunc: usageFuncNoDefaultValues,
Subcommands: []*ffcli.Command{
{
Name: "set",
Exec: runShareSet,
ShortHelp: "[ALPHA] set a share",
UsageFunc: usageFunc,
},
{
Name: "rename",
ShortHelp: "[ALPHA] rename a share",
Exec: runShareRename,
UsageFunc: usageFunc,
},
{
Name: "remove",
ShortHelp: "[ALPHA] remove a share",
Exec: runShareRemove,
UsageFunc: usageFunc,
},
{
Name: "list",
ShortHelp: "[ALPHA] list current shares",
Exec: runShareList,
UsageFunc: usageFunc,
},
},
Exec: func(context.Context, []string) error {
return errors.New("share subcommand required; run 'tailscale share -h' for details")
},
}
// runShareSet is the entry point for the "tailscale share set" command.
func runShareSet(ctx context.Context, args []string) error {
if len(args) != 2 {
return fmt.Errorf("usage: tailscale %v", shareSetUsage)
}
name, path := args[0], args[1]
err := localClient.TailFSShareSet(ctx, &tailfs.Share{
Name: name,
Path: path,
})
if err == nil {
fmt.Printf("Set share %q at %q\n", name, path)
}
return err
}
// runShareRemove is the entry point for the "tailscale share remove" command.
func runShareRemove(ctx context.Context, args []string) error {
if len(args) != 1 {
return fmt.Errorf("usage: tailscale %v", shareRemoveUsage)
}
name := args[0]
err := localClient.TailFSShareRemove(ctx, name)
if err == nil {
fmt.Printf("Removed share %q\n", name)
}
return err
}
// runShareRename is the entry point for the "tailscale share rename" command.
func runShareRename(ctx context.Context, args []string) error {
if len(args) != 2 {
return fmt.Errorf("usage: tailscale %v", shareRenameUsage)
}
oldName := args[0]
newName := args[1]
err := localClient.TailFSShareRename(ctx, oldName, newName)
if err == nil {
fmt.Printf("Renamed share %q to %q\n", oldName, newName)
}
return err
}
// runShareList is the entry point for the "tailscale share list" command.
func runShareList(ctx context.Context, args []string) error {
if len(args) != 0 {
return fmt.Errorf("usage: tailscale %v", shareListUsage)
}
shares, err := localClient.TailFSShareList(ctx)
if err != nil {
return err
}
longestName := 4 // "name"
longestPath := 4 // "path"
longestAs := 2 // "as"
for _, share := range shares {
if len(share.Name) > longestName {
longestName = len(share.Name)
}
if len(share.Path) > longestPath {
longestPath = len(share.Path)
}
if len(share.As) > longestAs {
longestAs = len(share.As)
}
}
formatString := fmt.Sprintf("%%-%ds %%-%ds %%s\n", longestName, longestPath)
fmt.Printf(formatString, "name", "path", "as")
fmt.Printf(formatString, strings.Repeat("-", longestName), strings.Repeat("-", longestPath), strings.Repeat("-", longestAs))
for _, share := range shares {
fmt.Printf(formatString, share.Name, share.Path, share.As)
}
return nil
}
func buildShareLongHelp() string {
longHelpAs := ""
if tailfs.AllowShareAs() {
longHelpAs = shareLongHelpAs
}
return fmt.Sprintf(shareLongHelpBase, longHelpAs)
}
var shareLongHelpBase = `Tailscale share allows you to share directories with other machines on your tailnet.
In order to share folders, your node needs to have the node attribute "tailfs:share".
In order to access shares, your node needs to have the node attribute "tailfs:access".
For example, to enable sharing and accessing shares for all member nodes:
"nodeAttrs": [
{
"target": ["autogroup:member"],
"attr": [
"tailfs:share",
"tailfs:access",
],
}]
Each share is identified by a name and points to a directory at a specific path. For example, to share the path /Users/me/Documents under the name "docs", you would run:
$ tailscale share set docs /Users/me/Documents
Note that the system forces share names to lowercase to avoid problems with clients that don't support case-sensitive filenames.
Share names may only contain the letters a-z, underscore _, parentheses (), or spaces. Leading and trailing spaces are omitted.
All Tailscale shares have a globally unique path consisting of the tailnet, the machine name and the share name. For example, if the above share was created on the machine "mylaptop" on the tailnet "mydomain.com", the share's path would be:
/mydomain.com/mylaptop/docs
In order to access this share, other machines on the tailnet can connect to the above path on a WebDAV server running at 100.100.100.100:8080, for example:
http://100.100.100.100:8080/mydomain.com/mylaptop/docs
Permissions to access shares are controlled via ACLs. For example, to give yourself read/write access and give the group "home" read-only access to the above share, use the below ACL grants:
"grants": [
{
"src": ["mylogin@domain.com"],
"dst": ["mylaptop's ip address"],
"app": {
"tailscale.com/cap/tailfs": [{
"shares": ["docs"],
"access": "rw"
}]
}
},
{
"src": ["group:home"],
"dst": ["mylaptop"],
"app": {
"tailscale.com/cap/tailfs": [{
"shares": ["docs"],
"access": "ro"
}]
}
}]
To categorically give yourself access to all your shares, you can use the below ACL grant:
"grants": [
{
"src": ["autogroup:member"],
"dst": ["autogroup:self"],
"app": {
"tailscale.com/cap/tailfs": [{
"shares": ["*"],
"access": "rw"
}]
}
}]
Whenever either you or anyone in the group "home" connects to the share, they connect as if they are using your local machine user. They'll be able to read the same files as your user and if they create files, those files will be owned by your user.%s
You can rename shares, for example you could rename the above share by running:
$ tailscale share rename docs newdocs
You can remove shares by name, for example you could remove the above share by running:
$ tailscale share remove newdocs
You can get a list of currently published shares by running:
$ tailscale share list`
var shareLongHelpAs = `
If you want a share to be accessed as a different user, you can use sudo to accomplish this. For example, to create the aforementioned share as "theuser", you could run:
$ sudo -u theuser tailscale share set docs /Users/theuser/Documents`

View File

@@ -48,8 +48,8 @@ The 'tailscale ssh' wrapper adds a few things:
}
func runSSH(ctx context.Context, args []string) error {
if runtime.GOOS == "darwin" && version.IsSandboxedMacOS() && !envknob.UseWIPCode() {
return errors.New("The 'tailscale ssh' subcommand is not available on sandboxed macOS builds.\nUse the regular 'ssh' client instead.")
if runtime.GOOS == "darwin" && version.IsMacAppStore() && !envknob.UseWIPCode() {
return errors.New("The 'tailscale ssh' subcommand is not available on macOS builds distributed through the App Store or TestFlight.\nInstall the Standalone variant of Tailscale (download it from https://pkgs.tailscale.com), or use the regular 'ssh' client instead.")
}
if len(args) == 0 {
return errors.New("usage: ssh [user@]<host>")
@@ -103,6 +103,7 @@ func runSSH(ctx context.Context, args []string) error {
"-o", fmt.Sprintf("UserKnownHostsFile %q", knownHostsFile),
"-o", "UpdateHostKeys no",
"-o", "StrictHostKeyChecking yes",
"-o", "CanonicalizeHostname no", // https://github.com/tailscale/tailscale/issues/10348
)
// TODO(bradfitz): nc is currently broken on macOS:

View File

@@ -5,6 +5,7 @@ package cli
import (
"bytes"
"cmp"
"context"
"encoding/json"
"errors"
@@ -23,7 +24,6 @@ import (
"tailscale.com/ipn"
"tailscale.com/ipn/ipnstate"
"tailscale.com/net/interfaces"
"tailscale.com/util/cmpx"
"tailscale.com/util/dnsname"
)
@@ -228,8 +228,8 @@ func runStatus(ctx context.Context, args []string) error {
}
Stdout.Write(buf.Bytes())
if locBasedExitNode {
println()
println("# To see the full list of exit nodes, including location-based exit nodes, run `tailscale exit-node list` \n")
outln()
printf("# To see the full list of exit nodes, including location-based exit nodes, run `tailscale exit-node list` \n")
}
if len(st.Health) > 0 {
outln()
@@ -316,7 +316,7 @@ func ownerLogin(st *ipnstate.Status, ps *ipnstate.PeerStatus) string {
// their netmap. We've historically (2021-01..2023-08) always shown the
// sharer's name in the UI. Perhaps we want to show both here? But the CLI's
// a bit space constrained.
uid := cmpx.Or(ps.AltSharerUserID, ps.UserID)
uid := cmp.Or(ps.AltSharerUserID, ps.UserID)
if uid.IsZero() {
return "-"
}

View File

@@ -652,6 +652,7 @@ func upWorthyWarning(s string) bool {
return strings.Contains(s, healthmsg.TailscaleSSHOnBut) ||
strings.Contains(s, healthmsg.WarnAcceptRoutesOff) ||
strings.Contains(s, healthmsg.LockedOut) ||
strings.Contains(s, healthmsg.WarnExitNodeUsage) ||
strings.Contains(strings.ToLower(s), "update available: ")
}
@@ -726,7 +727,7 @@ func init() {
func addPrefFlagMapping(flagName string, prefNames ...string) {
prefsOfFlag[flagName] = prefNames
prefType := reflect.TypeOf(ipn.Prefs{})
prefType := reflect.TypeFor[ipn.Prefs]()
for _, pref := range prefNames {
t := prefType
for _, name := range strings.Split(pref, ".") {

View File

@@ -20,7 +20,7 @@ import (
var updateCmd = &ffcli.Command{
Name: "update",
ShortUsage: "update",
ShortHelp: "[BETA] Update Tailscale to the latest/different version",
ShortHelp: "Update Tailscale to the latest/different version",
Exec: runUpdate,
FlagSet: (func() *flag.FlagSet {
fs := newFlagSet("update")

View File

@@ -4,6 +4,7 @@
package cli
import (
"cmp"
"context"
"crypto/tls"
_ "embed"
@@ -21,7 +22,6 @@ import (
"github.com/peterbourgon/ff/v3/ffcli"
"tailscale.com/client/web"
"tailscale.com/ipn"
"tailscale.com/util/cmpx"
)
var webCmd = &ffcli.Command{
@@ -42,15 +42,17 @@ Tailscale, as opposed to a CLI or a native app.
webf.StringVar(&webArgs.listen, "listen", "localhost:8088", "listen address; use port 0 for automatic")
webf.BoolVar(&webArgs.cgi, "cgi", false, "run as CGI script")
webf.StringVar(&webArgs.prefix, "prefix", "", "URL prefix added to requests (for cgi or reverse proxies)")
webf.BoolVar(&webArgs.readonly, "readonly", false, "run web UI in read-only mode")
return webf
})(),
Exec: runWeb,
}
var webArgs struct {
listen string
cgi bool
prefix string
listen string
cgi bool
prefix string
readonly bool
}
func tlsConfigFromEnvironment() *tls.Config {
@@ -94,20 +96,26 @@ func runWeb(ctx context.Context, args []string) error {
if prefs, err := localClient.GetPrefs(ctx); err == nil {
existingWebClient = prefs.RunWebClient
}
if !existingWebClient {
var startedManagementClient bool // we started the management client
if !existingWebClient && !webArgs.readonly {
// Also start full client in tailscaled.
log.Printf("starting tailscaled web client at %s:%d\n", selfIP.String(), web.ListenPort)
log.Printf("starting tailscaled web client at http://%s\n", netip.AddrPortFrom(selfIP, web.ListenPort))
if err := setRunWebClient(ctx, true); err != nil {
return fmt.Errorf("starting web client in tailscaled: %w", err)
}
startedManagementClient = true
}
webServer, err := web.NewServer(web.ServerOpts{
opts := web.ServerOpts{
Mode: web.LoginServerMode,
CGIMode: webArgs.cgi,
PathPrefix: webArgs.prefix,
LocalClient: &localClient,
})
}
if webArgs.readonly {
opts.Mode = web.ReadOnlyServerMode
}
webServer, err := web.NewServer(opts)
if err != nil {
log.Printf("tailscale.web: %v", err)
return err
@@ -117,10 +125,10 @@ func runWeb(ctx context.Context, args []string) error {
case <-ctx.Done():
// Shutdown the server.
webServer.Shutdown()
if !webArgs.cgi && !existingWebClient {
if !webArgs.cgi && startedManagementClient {
log.Println("stopping tailscaled web client")
// When not in cgi mode, shut down the tailscaled
// web client on cli termination.
// web client on cli termination if we started it.
if err := setRunWebClient(context.Background(), false); err != nil {
log.Printf("stopping tailscaled web client: %v", err)
}
@@ -160,5 +168,5 @@ func setRunWebClient(ctx context.Context, val bool) error {
// urlOfListenAddr parses a given listen address into a formatted URL
func urlOfListenAddr(addr string) string {
host, port, _ := net.SplitHostPort(addr)
return fmt.Sprintf("http://%s", net.JoinHostPort(cmpx.Or(host, "127.0.0.1"), port))
return fmt.Sprintf("http://%s", net.JoinHostPort(cmp.Or(host, "127.0.0.1"), port))
}

View File

@@ -2,12 +2,12 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
filippo.io/edwards25519 from github.com/hdevalence/ed25519consensus
filippo.io/edwards25519/field from filippo.io/edwards25519
W 💣 github.com/alexbrainman/sspi from github.com/alexbrainman/sspi/negotiate+
W 💣 github.com/alexbrainman/sspi from github.com/alexbrainman/sspi/internal/common+
W github.com/alexbrainman/sspi/internal/common from github.com/alexbrainman/sspi/negotiate
W 💣 github.com/alexbrainman/sspi/negotiate from tailscale.com/net/tshttpproxy
L github.com/coreos/go-iptables/iptables from tailscale.com/util/linuxfw
L github.com/coreos/go-systemd/v22/dbus from tailscale.com/clientupdate
W 💣 github.com/dblohm7/wingoes from tailscale.com/util/winutil/authenticode+
W 💣 github.com/dblohm7/wingoes from github.com/dblohm7/wingoes/pe+
W 💣 github.com/dblohm7/wingoes/pe from tailscale.com/util/winutil/authenticode
github.com/fxamacker/cbor/v2 from tailscale.com/tka
L 💣 github.com/godbus/dbus/v5 from github.com/coreos/go-systemd/v22/dbus
@@ -18,18 +18,17 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
L github.com/google/nftables/expr from github.com/google/nftables+
L github.com/google/nftables/internal/parseexprfunc from github.com/google/nftables+
L github.com/google/nftables/xt from github.com/google/nftables/expr+
github.com/google/uuid from tailscale.com/util/quarantine+
github.com/google/uuid from tailscale.com/clientupdate+
github.com/gorilla/csrf from tailscale.com/client/web
github.com/gorilla/securecookie from github.com/gorilla/csrf
github.com/hdevalence/ed25519consensus from tailscale.com/tka+
github.com/hdevalence/ed25519consensus from tailscale.com/clientupdate/distsign+
L github.com/josharian/native from github.com/mdlayher/netlink+
L 💣 github.com/jsimonetti/rtnetlink from tailscale.com/net/interfaces+
L github.com/jsimonetti/rtnetlink/internal/unix from github.com/jsimonetti/rtnetlink
github.com/kballard/go-shellquote from tailscale.com/cmd/tailscale/cli
github.com/klauspost/compress/flate from nhooyr.io/websocket
💣 github.com/mattn/go-colorable from tailscale.com/cmd/tailscale/cli
💣 github.com/mattn/go-isatty from github.com/mattn/go-colorable+
L 💣 github.com/mdlayher/netlink from github.com/jsimonetti/rtnetlink+
L 💣 github.com/mdlayher/netlink from github.com/google/nftables+
L 💣 github.com/mdlayher/netlink/nlenc from github.com/jsimonetti/rtnetlink+
L github.com/mdlayher/netlink/nltest from github.com/google/nftables
L 💣 github.com/mdlayher/socket from github.com/mdlayher/netlink
@@ -38,7 +37,6 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
github.com/peterbourgon/ff/v3 from github.com/peterbourgon/ff/v3/ffcli
github.com/peterbourgon/ff/v3/ffcli from tailscale.com/cmd/tailscale/cli
github.com/peterbourgon/ff/v3/internal from github.com/peterbourgon/ff/v3
github.com/pkg/errors from github.com/gorilla/csrf
github.com/skip2/go-qrcode from tailscale.com/cmd/tailscale/cli
github.com/skip2/go-qrcode/bitset from github.com/skip2/go-qrcode+
github.com/skip2/go-qrcode/reedsolomon from github.com/skip2/go-qrcode
@@ -60,23 +58,24 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
L 💣 github.com/vishvananda/netlink/nl from github.com/tailscale/netlink
L github.com/vishvananda/netns from github.com/tailscale/netlink+
github.com/x448/float16 from github.com/fxamacker/cbor/v2
💣 go4.org/mem from tailscale.com/derp+
go4.org/netipx from tailscale.com/wgengine/filter+
💣 go4.org/mem from tailscale.com/client/tailscale+
go4.org/netipx from tailscale.com/net/tsaddr+
W 💣 golang.zx2c4.com/wireguard/windows/tunnel/winipcfg from tailscale.com/net/interfaces+
gopkg.in/yaml.v2 from sigs.k8s.io/yaml
k8s.io/client-go/util/homedir from tailscale.com/cmd/tailscale/cli
nhooyr.io/websocket from tailscale.com/derp/derphttp+
nhooyr.io/websocket from tailscale.com/control/controlhttp+
nhooyr.io/websocket/internal/errd from nhooyr.io/websocket
nhooyr.io/websocket/internal/util from nhooyr.io/websocket
nhooyr.io/websocket/internal/xsync from nhooyr.io/websocket
sigs.k8s.io/yaml from tailscale.com/cmd/tailscale/cli
sigs.k8s.io/yaml/goyaml.v2 from sigs.k8s.io/yaml
software.sslmate.com/src/go-pkcs12 from tailscale.com/cmd/tailscale/cli
software.sslmate.com/src/go-pkcs12/internal/rc2 from software.sslmate.com/src/go-pkcs12
tailscale.com from tailscale.com/version
tailscale.com/atomicfile from tailscale.com/ipn+
tailscale.com/client/tailscale from tailscale.com/cmd/tailscale/cli+
tailscale.com/client/tailscale/apitype from tailscale.com/cmd/tailscale/cli+
tailscale.com/atomicfile from tailscale.com/cmd/tailscale/cli+
tailscale.com/client/tailscale from tailscale.com/client/web+
tailscale.com/client/tailscale/apitype from tailscale.com/client/tailscale+
tailscale.com/client/web from tailscale.com/cmd/tailscale/cli
tailscale.com/clientupdate from tailscale.com/cmd/tailscale/cli+
tailscale.com/clientupdate from tailscale.com/client/web+
tailscale.com/clientupdate/distsign from tailscale.com/clientupdate
tailscale.com/cmd/tailscale/cli from tailscale.com/cmd/tailscale
tailscale.com/control/controlbase from tailscale.com/control/controlhttp
@@ -85,54 +84,55 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
tailscale.com/derp from tailscale.com/derp/derphttp
tailscale.com/derp/derphttp from tailscale.com/net/netcheck
tailscale.com/disco from tailscale.com/derp
tailscale.com/envknob from tailscale.com/cmd/tailscale/cli+
tailscale.com/envknob from tailscale.com/client/tailscale+
tailscale.com/health from tailscale.com/net/tlsdial
tailscale.com/health/healthmsg from tailscale.com/cmd/tailscale/cli
tailscale.com/hostinfo from tailscale.com/net/interfaces+
tailscale.com/ipn from tailscale.com/cmd/tailscale/cli+
tailscale.com/ipn/ipnstate from tailscale.com/cmd/tailscale/cli+
tailscale.com/licenses from tailscale.com/cmd/tailscale/cli+
tailscale.com/hostinfo from tailscale.com/client/web+
tailscale.com/ipn from tailscale.com/client/tailscale+
tailscale.com/ipn/ipnstate from tailscale.com/client/tailscale+
tailscale.com/licenses from tailscale.com/client/web+
tailscale.com/metrics from tailscale.com/derp
tailscale.com/net/dns/recursive from tailscale.com/net/dnsfallback
tailscale.com/net/dnscache from tailscale.com/derp/derphttp+
tailscale.com/net/dnscache from tailscale.com/control/controlhttp+
tailscale.com/net/dnsfallback from tailscale.com/control/controlhttp
tailscale.com/net/flowtrack from tailscale.com/wgengine/filter+
tailscale.com/net/flowtrack from tailscale.com/net/packet+
💣 tailscale.com/net/interfaces from tailscale.com/cmd/tailscale/cli+
tailscale.com/net/netaddr from tailscale.com/ipn+
tailscale.com/net/netcheck from tailscale.com/cmd/tailscale/cli
tailscale.com/net/neterror from tailscale.com/net/netcheck+
tailscale.com/net/netknob from tailscale.com/net/netns
tailscale.com/net/netmon from tailscale.com/net/sockstats+
tailscale.com/net/netmon from tailscale.com/cmd/tailscale/cli+
tailscale.com/net/netns from tailscale.com/derp/derphttp+
tailscale.com/net/netutil from tailscale.com/client/tailscale+
tailscale.com/net/packet from tailscale.com/wgengine/filter+
tailscale.com/net/packet from tailscale.com/wgengine/capture+
tailscale.com/net/ping from tailscale.com/net/netcheck
tailscale.com/net/portmapper from tailscale.com/net/netcheck+
tailscale.com/net/portmapper from tailscale.com/cmd/tailscale/cli+
tailscale.com/net/sockstats from tailscale.com/control/controlhttp+
tailscale.com/net/stun from tailscale.com/net/netcheck
L tailscale.com/net/tcpinfo from tailscale.com/derp
tailscale.com/net/tlsdial from tailscale.com/derp/derphttp+
tailscale.com/net/tsaddr from tailscale.com/net/interfaces+
💣 tailscale.com/net/tshttpproxy from tailscale.com/derp/derphttp+
tailscale.com/net/tlsdial from tailscale.com/cmd/tailscale/cli+
tailscale.com/net/tsaddr from tailscale.com/client/web+
💣 tailscale.com/net/tshttpproxy from tailscale.com/clientupdate/distsign+
tailscale.com/net/wsconn from tailscale.com/control/controlhttp+
tailscale.com/paths from tailscale.com/cmd/tailscale/cli+
💣 tailscale.com/safesocket from tailscale.com/cmd/tailscale/cli+
tailscale.com/syncs from tailscale.com/net/netcheck+
tailscale.com/tailcfg from tailscale.com/cmd/tailscale/cli+
tailscale.com/paths from tailscale.com/client/tailscale+
💣 tailscale.com/safesocket from tailscale.com/client/tailscale+
tailscale.com/syncs from tailscale.com/cmd/tailscale/cli+
tailscale.com/tailcfg from tailscale.com/client/tailscale+
tailscale.com/tailfs from tailscale.com/cmd/tailscale/cli+
tailscale.com/tka from tailscale.com/client/tailscale+
W tailscale.com/tsconst from tailscale.com/net/interfaces
tailscale.com/tstime from tailscale.com/control/controlhttp+
tailscale.com/tstime/mono from tailscale.com/tstime/rate
tailscale.com/tstime/rate from tailscale.com/wgengine/filter+
tailscale.com/tstime/rate from tailscale.com/cmd/tailscale/cli+
tailscale.com/types/dnstype from tailscale.com/tailcfg
tailscale.com/types/empty from tailscale.com/ipn
tailscale.com/types/ipproto from tailscale.com/net/flowtrack+
tailscale.com/types/key from tailscale.com/derp+
tailscale.com/types/lazy from tailscale.com/version+
tailscale.com/types/logger from tailscale.com/cmd/tailscale/cli+
tailscale.com/types/key from tailscale.com/client/tailscale+
tailscale.com/types/lazy from tailscale.com/util/testenv+
tailscale.com/types/logger from tailscale.com/client/web+
tailscale.com/types/netmap from tailscale.com/ipn
tailscale.com/types/nettype from tailscale.com/net/netcheck+
tailscale.com/types/opt from tailscale.com/net/netcheck+
tailscale.com/types/opt from tailscale.com/client/tailscale+
tailscale.com/types/persist from tailscale.com/ipn
tailscale.com/types/preftype from tailscale.com/cmd/tailscale/cli+
tailscale.com/types/ptr from tailscale.com/hostinfo+
@@ -142,34 +142,34 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
tailscale.com/util/clientmetric from tailscale.com/net/netcheck+
tailscale.com/util/cloudenv from tailscale.com/net/dnscache+
tailscale.com/util/cmpver from tailscale.com/net/tshttpproxy+
tailscale.com/util/cmpx from tailscale.com/cmd/tailscale/cli+
tailscale.com/util/ctxkey from tailscale.com/types/logger
L 💣 tailscale.com/util/dirwalk from tailscale.com/metrics
tailscale.com/util/dnsname from tailscale.com/cmd/tailscale/cli+
tailscale.com/util/groupmember from tailscale.com/client/web
tailscale.com/util/httpm from tailscale.com/client/tailscale+
tailscale.com/util/lineread from tailscale.com/net/interfaces+
tailscale.com/util/lineread from tailscale.com/hostinfo+
L tailscale.com/util/linuxfw from tailscale.com/net/netns
tailscale.com/util/mak from tailscale.com/net/netcheck+
tailscale.com/util/mak from tailscale.com/cmd/tailscale/cli+
tailscale.com/util/multierr from tailscale.com/control/controlhttp+
tailscale.com/util/must from tailscale.com/cmd/tailscale/cli+
tailscale.com/util/must from tailscale.com/clientupdate/distsign+
tailscale.com/util/nocasemaps from tailscale.com/types/ipproto
tailscale.com/util/quarantine from tailscale.com/cmd/tailscale/cli
tailscale.com/util/set from tailscale.com/health+
tailscale.com/util/set from tailscale.com/derp+
tailscale.com/util/singleflight from tailscale.com/net/dnscache+
tailscale.com/util/slicesx from tailscale.com/net/dnscache+
tailscale.com/util/slicesx from tailscale.com/net/dns/recursive+
tailscale.com/util/syspolicy from tailscale.com/ipn
tailscale.com/util/testenv from tailscale.com/cmd/tailscale/cli
tailscale.com/util/truncate from tailscale.com/cmd/tailscale/cli
tailscale.com/util/vizerror from tailscale.com/types/ipproto+
💣 tailscale.com/util/winutil from tailscale.com/hostinfo+
tailscale.com/util/vizerror from tailscale.com/tailcfg+
💣 tailscale.com/util/winutil from tailscale.com/clientupdate+
W 💣 tailscale.com/util/winutil/authenticode from tailscale.com/clientupdate
tailscale.com/version from tailscale.com/cmd/tailscale/cli+
tailscale.com/version/distro from tailscale.com/cmd/tailscale/cli+
tailscale.com/version from tailscale.com/client/web+
tailscale.com/version/distro from tailscale.com/client/web+
tailscale.com/wgengine/capture from tailscale.com/cmd/tailscale/cli
tailscale.com/wgengine/filter from tailscale.com/types/netmap
golang.org/x/crypto/argon2 from tailscale.com/tka
golang.org/x/crypto/blake2b from golang.org/x/crypto/nacl/box+
golang.org/x/crypto/blake2s from tailscale.com/control/controlbase+
golang.org/x/crypto/blake2b from golang.org/x/crypto/argon2+
golang.org/x/crypto/blake2s from tailscale.com/clientupdate/distsign+
golang.org/x/crypto/chacha20 from golang.org/x/crypto/chacha20poly1305
golang.org/x/crypto/chacha20poly1305 from crypto/tls+
golang.org/x/crypto/cryptobyte from crypto/ecdsa+
@@ -189,18 +189,18 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
golang.org/x/net/http2/hpack from net/http
golang.org/x/net/icmp from tailscale.com/net/ping
golang.org/x/net/idna from golang.org/x/net/http/httpguts+
golang.org/x/net/ipv4 from golang.org/x/net/icmp+
golang.org/x/net/ipv6 from golang.org/x/net/icmp+
golang.org/x/net/ipv4 from github.com/miekg/dns+
golang.org/x/net/ipv6 from github.com/miekg/dns+
golang.org/x/net/proxy from tailscale.com/net/netns
D golang.org/x/net/route from net+
golang.org/x/oauth2 from golang.org/x/oauth2/clientcredentials
golang.org/x/oauth2/clientcredentials from tailscale.com/cmd/tailscale/cli
golang.org/x/oauth2/internal from golang.org/x/oauth2+
golang.org/x/sync/errgroup from tailscale.com/derp+
golang.org/x/sys/cpu from golang.org/x/crypto/blake2b+
LD golang.org/x/sys/unix from tailscale.com/net/netns+
W golang.org/x/sys/windows from golang.org/x/sys/windows/registry+
W golang.org/x/sys/windows/registry from golang.zx2c4.com/wireguard/windows/tunnel/winipcfg+
golang.org/x/sync/errgroup from github.com/mdlayher/socket+
golang.org/x/sys/cpu from github.com/josharian/native+
LD golang.org/x/sys/unix from github.com/google/nftables+
W golang.org/x/sys/windows from github.com/dblohm7/wingoes+
W golang.org/x/sys/windows/registry from github.com/dblohm7/wingoes+
W golang.org/x/sys/windows/svc from golang.org/x/sys/windows/svc/mgr+
W golang.org/x/sys/windows/svc/mgr from tailscale.com/util/winutil
golang.org/x/text/secure/bidirule from golang.org/x/net/idna
@@ -210,14 +210,14 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
golang.org/x/time/rate from tailscale.com/cmd/tailscale/cli+
archive/tar from tailscale.com/clientupdate
bufio from compress/flate+
bytes from bufio+
bytes from archive/tar+
cmp from slices+
compress/flate from compress/gzip+
compress/gzip from net/http+
compress/zlib from image/png+
compress/zlib from debug/pe+
container/list from crypto/tls+
context from crypto/tls+
crypto from crypto/ecdsa+
crypto from crypto/ecdh+
crypto/aes from crypto/ecdsa+
crypto/cipher from crypto/aes+
crypto/des from crypto/tls+
@@ -235,16 +235,16 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
crypto/sha256 from crypto/tls+
crypto/sha512 from crypto/ecdsa+
crypto/subtle from crypto/aes+
crypto/tls from github.com/tcnksm/go-httpstat+
crypto/tls from github.com/miekg/dns+
crypto/x509 from crypto/tls+
crypto/x509/pkix from crypto/x509+
database/sql/driver from github.com/google/uuid
W debug/dwarf from debug/pe
W debug/pe from github.com/dblohm7/wingoes/pe
embed from tailscale.com/cmd/tailscale/cli+
encoding from encoding/json+
embed from crypto/internal/nistec+
encoding from encoding/gob+
encoding/asn1 from crypto/x509+
encoding/base32 from tailscale.com/tka+
encoding/base32 from github.com/fxamacker/cbor/v2+
encoding/base64 from encoding/json+
encoding/binary from compress/gzip+
encoding/gob from github.com/gorilla/securecookie
@@ -252,64 +252,64 @@ tailscale.com/cmd/tailscale dependencies: (generated by github.com/tailscale/dep
encoding/json from expvar+
encoding/pem from crypto/tls+
encoding/xml from github.com/tailscale/goupnp+
errors from bufio+
errors from archive/tar+
expvar from tailscale.com/derp+
flag from github.com/peterbourgon/ff/v3+
fmt from compress/flate+
hash from crypto+
fmt from archive/tar+
hash from compress/zlib+
hash/adler32 from compress/zlib
hash/crc32 from compress/gzip+
hash/maphash from go4.org/mem
html from tailscale.com/ipn/ipnstate+
html from html/template+
html/template from github.com/gorilla/csrf
image from github.com/skip2/go-qrcode+
image/color from github.com/skip2/go-qrcode+
image/png from github.com/skip2/go-qrcode
io from bufio+
io/fs from crypto/x509+
io/ioutil from golang.org/x/sys/cpu+
io from archive/tar+
io/fs from archive/tar+
io/ioutil from github.com/godbus/dbus/v5+
log from expvar+
log/internal from log
maps from tailscale.com/types/views+
math from compress/flate+
maps from tailscale.com/clientupdate+
math from archive/tar+
math/big from crypto/dsa+
math/bits from compress/flate+
math/rand from math/big+
mime from mime/multipart+
math/rand from github.com/mdlayher/netlink+
mime from golang.org/x/oauth2/internal+
mime/multipart from net/http
mime/quotedprintable from mime/multipart
net from crypto/tls+
net/http from expvar+
net/http/cgi from tailscale.com/cmd/tailscale/cli
net/http/httptrace from github.com/tcnksm/go-httpstat+
net/http/httputil from tailscale.com/cmd/tailscale/cli+
net/http/httputil from tailscale.com/client/web+
net/http/internal from net/http+
net/netip from net+
net/netip from go4.org/netipx+
net/textproto from golang.org/x/net/http/httpguts+
net/url from crypto/x509+
os from crypto/rand+
os/exec from github.com/toqueteos/webbrowser+
os/exec from github.com/coreos/go-iptables/iptables+
os/signal from tailscale.com/cmd/tailscale/cli
os/user from tailscale.com/util/groupmember+
path from html/template+
path/filepath from crypto/x509+
reflect from crypto/x509+
regexp from github.com/tailscale/goupnp/httpu+
os/user from archive/tar+
path from archive/tar+
path/filepath from archive/tar+
reflect from archive/tar+
regexp from github.com/coreos/go-iptables/iptables+
regexp/syntax from regexp
runtime/debug from tailscale.com/util/singleflight+
runtime/debug from nhooyr.io/websocket/internal/xsync+
runtime/trace from testing
slices from tailscale.com/cmd/tailscale/cli+
sort from compress/flate+
strconv from compress/flate+
strings from bufio+
sync from compress/flate+
slices from tailscale.com/client/web+
sort from archive/tar+
strconv from archive/tar+
strings from archive/tar+
sync from archive/tar+
sync/atomic from context+
syscall from crypto/rand+
syscall from archive/tar+
testing from tailscale.com/util/syspolicy
text/tabwriter from github.com/peterbourgon/ff/v3/ffcli+
text/template from html/template
text/template/parse from html/template+
time from compress/gzip+
time from archive/tar+
unicode from bytes+
unicode/utf16 from encoding/asn1+
unicode/utf16 from crypto/x509+
unicode/utf8 from bufio+

Some files were not shown because too many files have changed in this diff Show More