Page MenuHomePhabricator

PostBuild changing the status of successful builds to failure for no apparent reason
Open, Needs TriagePublic

Description

Over the last couple weeks or so, I'm noticing relatively many instances where the post-build steps will change the status of an otherwise successful build to failure, for no apparent reason. Some examples:

They're all like:

+ exec docker run --entrypoint=bash --volume /srv/jenkins/workspace/quibble-apitests-only-vendor-php83/src:/src --volume /srv/jenkins/workspace/quibble-apitests-only-vendor-php83/cache:/cache --volume /srv/jenkins/workspace/quibble-apitests-only-vendor-php83/log:/log --security-opt seccomp=unconfined --init --rm --label jenkins.job=quibble-apitests-only-vendor-php83 --label jenkins.build=12947 --env-file /dev/fd/63 docker-registry.wikimedia.org/releng/quibble-bullseye-php83:1.16.0 -c '([[ "$ZUUL_PIPELINE" =~ ^(gate-and-submit|postmerge) ]] && npm cache verify) || :'
++ set +x
[PostBuildScript] - [INFO] Executing post build scripts.
Waiting for the completion of castor-save-workspace-cache
Build step 'Execute scripts' changed build result to FAILURE
Build step 'Execute scripts' marked build as failure

Event Timeline

I'm pretty sure this has a dupe somewhere, but yes, it seems to have become somewhat worse of late.

Oh yeah, maybe T409479: quibble-with-gated-extensions-vendor-mysql-php81 failure due to postbuild jobs (which might want a slightly more general title if this is merged into it).
There's also T352319: castor-save-workspace-cache aborted during postbuild, which seems like it might be describing a similar outcome (an apparent failure of castor-save-workspace-cache), but maybe with different symptoms beforehand (which suggest that it might be a different problem).