[
https://issues.apache.org/jira/browse/OOZIE-3615?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sivakumar Ramaswamy updated OOZIE-3615:
---------------------------------------
Description:
Hello,
oozie workflow runs forever when OS assign oozie job PID to kworker .
Below returns true always:
ssh -o PasswordAuthentication=no -o KbdInteractiveDevices=no -o
StrictHostKeyChecking=no -o ConnectTimeout=20 hadoop@localhost ps -p <PID>
Reproduce:
hard to reproduce and there are 1500 jobs runs at a time . This issue happen
sporadically.
tried by changing below value but , no luck.
oozie.service.ActionCheckerService.action.check.delay
But no luck.
solution:
it would be great if we can check pid and process name as well.
was:
Hello,
oozie workflow runs forever when OS assign PID to kworker .
Below returns true always:
ssh -o PasswordAuthentication=no -o KbdInteractiveDevices=no -o
StrictHostKeyChecking=no -o ConnectTimeout=20 hadoop@localhost ps -p <PID>
Reproduce:
hard to reproduce and there are 1500 jobs runs at a time . This issue happen
sporadically.
Reduced
oozie.service.ActionCheckerService.action.check.delay
But no luck.
solution:
it would be great if we can check pid and process name as well.
> sshaction runs infinitely as OS assign same PID to kworker
> ----------------------------------------------------------
>
> Key: OOZIE-3615
> URL: https://issues.apache.org/jira/browse/OOZIE-3615
> Project: Oozie
> Issue Type: Bug
> Components: scripts
> Affects Versions: 5.1.0, 5.2.0
> Reporter: Sivakumar Ramaswamy
> Priority: Major
>
> Hello,
> oozie workflow runs forever when OS assign oozie job PID to kworker .
> Below returns true always:
> ssh -o PasswordAuthentication=no -o KbdInteractiveDevices=no -o
> StrictHostKeyChecking=no -o ConnectTimeout=20 hadoop@localhost ps -p <PID>
>
> Reproduce:
> hard to reproduce and there are 1500 jobs runs at a time . This issue happen
> sporadically.
>
> tried by changing below value but , no luck.
> oozie.service.ActionCheckerService.action.check.delay
>
> But no luck.
>
> solution:
> it would be great if we can check pid and process name as well.
>
>
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
