Details
-
Bug
-
Status: Closed
-
Minor
-
Resolution: Fixed
-
None
-
None
Description
https://issues.apache.org/jira/browse/YUNIKORN-657 introduced a change that set the pods to failed status when there's an application scheduling error. But the update of pod status may occasionally fail due to conflicts (e.g. the pods have been modified elsewhere). Need to get the latest pod before applying the status update. Also retry with exponential backoff if conflicts still exist
Attachments
Issue Links
- is caused by
-
YUNIKORN-657 Expose reason of application failure to pods
- Closed