Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
Description
Currently when a compactor worker thread is stuck, or working too long on a compaction the the initiator might decide to start a new compaction because of a timeout, but old worker might still wait for the results of the job.
It would be good to cancel the worker as well after the timeout is reached
Attachments
Attachments
Issue Links
- causes
-
HIVE-23591 When Worker fails to connect to metastore it should wait before retrying
- Closed
- links to