[
https://issues.apache.org/jira/browse/MAPREDUCE-3186?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eric Payne updated MAPREDUCE-3186:
----------------------------------
Attachment: MAPREDUCE-3186.v2.txt
I have addressed all of the following issues except for the unit tests. I
wanted to get the new patch up here so you could look at it while I
simulatneously address the unit tests.
Mahadev wrote:
> We should probably avoid reading the config entry everytime we call get
> resources. The maxretry can be inited in init() call.
Done. I initialzed variables in the init.
> Regarding the test, you should be able to mock failure the communicate with
> the RM and make sure that an internal error is generated. Also, if the MRApp
> shutsdown on an internal error.
Still working on the unit tests.
Sid wrote:
> can the timeout be changed from #failedAttempts to actual time. Currently the
> MRAM sends a heartbeat at a specific interval (2s default). This may change
> to ask for containers the moment a task needs to be launched.
Changed to a timed interval that will keep trying to contact the RM and error
out only if a certain amount of time has expired.
> The LocalContainerAllocator (used by uberized jobs) would need to be changed
> as well.
I have changed the LocalContainerAllocator code as well. Good catch.
> User jobs are getting hanged if the Resource manager process goes down and
> comes up while job is getting executed.
> ------------------------------------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-3186
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3186
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Environment: linux
> Reporter: Ramgopal N
> Assignee: Eric Payne
> Priority: Blocker
> Labels: test
> Attachments: MAPREDUCE-3186.v1.txt, MAPREDUCE-3186.v2.txt
>
>
> If the resource manager is restarted while the job execution is in progress,
> the job is getting hanged.
> UI shows the job as running.
> In the RM log, it is throwing an error "ERROR
> org.apache.hadoop.yarn.server.resourcemanager.ApplicationMasterService:
> AppAttemptId doesnt exist in cache appattempt_1318579738195_0004_000001"
> In the console MRAppMaster and Runjar processes are not getting killed
--
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
