[
https://issues.apache.org/jira/browse/MAPREDUCE-3614?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vinod Kumar Vavilapalli updated MAPREDUCE-3614:
-----------------------------------------------
Fix Version/s: 0.23.2
Target Version/s: 0.23.2 (was: 0.23.1)
Status: Open (was: Patch Available)
- FindBugs warning is still not fixed as Jenkins reports.
- Not sure why you need all the magic with the FileSystem objects, can you
explain?
- You can make {{MRAppMaster.MRAppMasterShutdownHook}} extend
{{CompositeServiceShutdownHook}}
- RMCommunicator: It seems unnatural to set killed final state if the job was
running. I think you should set a flag with RMCommunicator similar to
JobHistoryEventHandler.
JobHistoryEventHandler (JHEH):
- Trivial: {{isSigTermed == false}} can be written as {{!isSigTermed}}
- Again trivial: The sigterm flag will never go back to false from true. So
the code snippet "while (isSigTermed && jobIt.hasNext()) .." can be rewritten
as "if (isSigTermed) { while .. }"
- For the same snippet, can you simply add a comment saying, we are doing this
so as to drain everything and only write finish-events to shutdown as fast as
possible.
The test seems very wired into the code. Can we instead do something like this:
- Send 4 history events to JHEH and properly shutdown it and verify that 4
events are logged.
- Send 4 history events to JHEH and shut it down via SIGTERM and verify that
only 1 job-finish event is logged.
bq. Is there a compelling need to maintain a parallel set of JobIds? Why not
check the fileMap values to see if the writer is open?
+1.
> finalState UNDEFINED if AM is killed by hand
> ---------------------------------------------
>
> Key: MAPREDUCE-3614
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3614
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Ravi Prakash
> Assignee: Ravi Prakash
> Fix For: 0.23.2
>
> Attachments: MAPREDUCE-3614.branch-0.23.patch, MAPREDUCE-3614.patch,
> MAPREDUCE-3614.patch, MAPREDUCE-3614.patch, MAPREDUCE-3614.patch
>
>
> Courtesy [~dcapwell]
> {quote}
> If the AM is running and you kill the process (sudo kill #pid), the State in
> Yarn would be FINISHED and FinalStatus is UNDEFINED. The Tracking UI would
> say "History" and point to the proxy url (which will redirect to the history
> server).
> The state should be more descriptive that the job failed and the tracker url
> shouldn't point to the history server.
> {quote}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
