[
https://issues.apache.org/jira/browse/OOZIE-1507?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rohini Palaniswamy updated OOZIE-1507:
--------------------------------------
Attachment: OOZIE-1507-2.patch
Good catch. Uploaded patch with that fix.
As far to why lock is not released, a command gets stuck in call to Oracle
database in loadState till restart while retrieving contents of blob in
Workflow Job table.
https://forums.oracle.com/message/9524124#9524124
We are seeing this only for a very big workflow with 80 forks where the blob
size is huge. This happens when there are lot of calls at the same time trying
to load the same workflow. That is the case during kill and one or two threads
hang.
> Command queue filling up with duplicate commands from RecoveryService
> ---------------------------------------------------------------------
>
> Key: OOZIE-1507
> URL: https://issues.apache.org/jira/browse/OOZIE-1507
> Project: Oozie
> Issue Type: Bug
> Reporter: Rohini Palaniswamy
> Assignee: Rohini Palaniswamy
> Fix For: 4.0.0
>
> Attachments: OOZIE-1507-1.patch, OOZIE-1507-2.patch
>
>
> For some reason one of the commands does not release the lock, which we
> have not been able to figure out why. Duplicate commands from RecoveryService
> fill up the queue and bring the server down to a halt. Couple of commands
> have been fixed in OOZIE-1473 and OOZIE-1433. But there are more.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
