Failing to pull docker image gitlab/gitlab-runner-helper:x86_64-1644837a ...
Summary
jobs are failing after 1hour, with:
[INFO] resource stream source: context loader classpath
Pulling docker image gitlab/gitlab-runner-helper:x86_64-1644837a ...
ERROR: Job failed: execution took longer than 1h0m0s seconds
Steps to reproduce
retry any of
- https://gitlab.com/apache-isis-committers/apache-isis/-/jobs/92476933
- https://gitlab.com/apache-isis-committers/apache-isis/-/jobs/92245893
- https://gitlab.com/apache-isis-committers/apache-isis/-/jobs/92196682
previous build had succeeded:
https://gitlab.com/apache-isis-committers/apache-isis/-/jobs/91474720
difference in behaviour is approx at line 6816 in the log file, where the failing builds show
section_end:1535326315:build_script
section_start:1535326315:after_script
section_end:1535326316:after_script
section_start:1535326316:upload_artifacts_on_failure
Pulling docker image gitlab/gitlab-runner-helper:x86_64-1644837a ...
section_end:1535326316:upload_artifacts_on_failure
ERROR: Job failed: execution took longer than 1h0m0s seconds
Actual behavior
It looks like the CI job that fails was running a test which then ended unexpectedly. The upload_artifacts_on_failure script then failed, attempting to use a docker image that couldn't be found.
Expected behavior
job shouldn't be failing.
What is going on
https://gitlab.com/gitlab-org/gitlab-ee/issues/10727#note_160094520
Proposal
- Don't run the artifact push when there is no artifacts to upload
- Add more verbose logging about artifacts, which might be covered with #4375
Edited by Steve Xuereb