Repository navigation
fix(cluster): run the join wait from the image's filesystem - #124
Merged
Merged
Conversation
Docker mounts a service's tmpfs noexec, and k3s finds tailscale with Go's exec.LookPath, which passes over a file it may not execute, so the wrapper under /run was never run and a joining node on Tailscale's control plane still never joined. The wrapper now lives under /usr/local/lib/github-runner-node/bin, is removed again when the node joins nothing, and logs a line to the container's log after its wait, which the mesh integration test now requires of every joining server. Fixes #116
grep -q stops at the first match, and under pipefail the docker logs it cut off with a broken pipe failed the check even when the line was there.
Mearman
marked this pull request as ready for review
October 20, 2026 21:36
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
This security review couldn’t start because the paying account has exhausted its Codex allowance and has insufficient available credits. The payer can check their balance and add credits in the Codex usage dashboard, then try again. If you do not manage the paying account, contact this repository's admins. |
|
🎉 This PR is included in version 2.7.1 🎉 The release is available on:
Your semantic-release bot 📦🚀 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #116, which #117 did not: the wrapper it writes into
/run/github-runner-node/binis never run, because Docker mounts the service's/runtmpfsnoexec(tmpfs /run tmpfs rw,nosuid,nodev,noexecin a running node) and k3s lookstailscaleup with Go'sexec.LookPath, which passes over a file it may not execute. On a fresh three-server cluster on Tailscale's control plane built from v2.5.0, the joining servers kept failing with "no such host" with the wrapper in place.The wrapper now lives in
/usr/local/lib/github-runner-node/binon the image's filesystem, still only on k3s'sPATH. That filesystem outlives a restart, so the entrypoint rewrites the wrapper at every start and removes it when the node joins nothing. After its wait it writesgithub-runner-node: <host> is a peer and resolves after tailscale upto the container's log (PID 1's standard error, best effort). With this entrypoint on the same failing server, the line appeared in its log and the "no such host" failures stopped.tests/mesh/run.shnow fails unless every joining server logged that line, so the integration test shows k3s really ran the wrapper in a container, whichever control server it uses; on v2.5.0 it would fail. The unit tests now keep the wrapper's directory clear of every tmpfs and volume either service mounts, and cover the stale wrapper's removal and the log line. Each was broken in turn and a test failed for each.