Actions
action #76765
openjob is incomplete with reason just being "api failure" and no logs can be uploaded due to OOM condition on worker, improve reason to point to potential causes
Start date:
Due date:
% Done:
0%
Estimated time:
Description
Motivation¶
See the support case #75454 . In some cases jobs can incomplete with reason just being "api failure" with no more information. Also no logs can be uploaded so it is hard for the test reviewer to find out what happened. We could improve the reason to point to potential causes, like in this case OOM on the worker.
Acceptance criteria¶
- AC1: No bare reason "api failure" but a reason with hints what to check or what could be potential error sources to dig into further
Suggestions¶
- Investigate the code searching for the search term
api failure
and try to find out what reasons could lead to the worker returning just "api failure" without further details. - Extend the reason string or distinguish further what could have gone wrong here
Workaround¶
If anyone encounters incomplete jobs with just "api failure", check the local worker system journal for errors
Updated by okurz about 4 years ago
- Copied from action #75454: sometimes clone job is incomplete because of api failure added
Actions