Repository navigation
fix: reconcile Linux zombie runners in the same PID namespace - #2606
Merged
Merged
Conversation
Contributor
 |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
quifox
changed the title
nicobailon
force-pushed
the
fix/zombie-pid-liveness
branch
from
October 1, 2026 06:29
bd22b33 to
a7862e4
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
On Linux, kill(pid, 0) succeeds for a zombie runner, leaving a recent run marked running until the stale threshold. After a successful PID probe, inspect /proc//stat for state Z only when the recorded PID namespace and the observer's namespace are both present and equal. Runs with missing or mismatched namespace identity retain the existing stale threshold. Procfs read failures retain the signal-probe fallback, and EPERM remains unknown.
The generic liveness helper keeps signal-only probing by default. Reconciliation and workflow reattachment explicitly enable zombie detection for verified namespaces, so an exited runner without a result cannot keep a reattached workflow waiting. Unscoped PID checks retain their existing behavior.
The regressions create real Linux zombies whose comm contains a closing parenthesis. They check immediate repair and read-only reattachment exit detection for matching namespaces, conservative handling for missing/mismatched identities, stale repair for uncertain reconciliation cases, and stable completion results. Missing/mismatched reattachment cases still accept a later published result. Namespace variations use controlled status/observer fixtures; the tests do not launch a live runner in a separate namespace. The Unreleased changelog credits @quifox and #2606.
Validation on Linux with Node v24.19.0:
CI and automated review must be checked on the latest PR head.