Prevent auto-retry AccessDenied error from stopping transition to failed
What does this MR do?
Fixes a bug, when Gitlab::Access::AccessDeniedError
raised in Ci::RetryBuildService
prevents job transition to failed
state.
Are there points in the code the reviewer needs to double check?
Why was this MR needed?
If a job is configured to be auto-retried, then job.drop!
automatically calls the .retry()
method. Usually this will be OK. But if there happened something that disallows user to retry a particular job (e.g. user lost access to the project) since the original job was started, Ci::RetryBuildService
will raise a Gitlab::Access::AccessDeniedError
exception.
It becomes a problem, when such job is planned to be dropped by StuckCiJobsWorker
. With this worker we're looking for jobs that we treat as stuck
and we try to drop them in batches. If at least one of the jobs is configured to be auto-retried, and it's in a state that causes Gitlab::Access::AccessDeniedError
to be raised from retry service, the whole batch will be rolled-back by the DB transaction. This eventually prevents StuckCiJobsWorker
from cleaning stale and stuck jobs.
Pleas look into gitlab-com/infrastructure#3866 for a live example of such situation that happened on GitLab.com
Screenshots (if relevant)
Does this MR meet the acceptance criteria?
-
Changelog entry added, if necessary -
Documentation created/updated -
API support added -
Tests added for this feature/bug - Review
-
Has been reviewed by UX -
Has been reviewed by Frontend -
Has been reviewed by Backend -
Has been reviewed by Database
-
-
Conform by the merge request performance guides -
Conform by the style guides -
Squashed related commits together -
Internationalization required/considered -
End-to-end tests pass ( package-and-qa
manual pipeline job)