[ 
https://issues.apache.org/jira/browse/YARN-9336?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16807607#comment-16807607
 ] 

Tarun Parimi commented on YARN-9336:
------------------------------------

Hi [~rohithsharma] ,

I have attached a branch-2.9 patch for backport . Looks like a jenkins issue in 
branch-2.9 is causing the build to fail. Local build for hadoop branch-2.9 runs 
fine with this patch. I have also manually verified the fix in my local 
cluster. The patch contains the same change as the one you already committed in 
trunk. Please review and commit to branch-2.9 when you get time.

> JobHistoryServer leaks CLOSE_WAIT tcp connections when using 
> LogAggregationIndexedFileController
> ------------------------------------------------------------------------------------------------
>
>                 Key: YARN-9336
>                 URL: https://issues.apache.org/jira/browse/YARN-9336
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: log-aggregation
>    Affects Versions: 2.9.0
>            Reporter: Tarun Parimi
>            Assignee: Tarun Parimi
>            Priority: Major
>             Fix For: 3.3.0
>
>         Attachments: YARN-9336-branch-2.9.001.patch, 
> YARN-9336-branch-2.9.002.patch, YARN-9336-branch-2.9.003.patch, 
> YARN-9336.001.patch, YARN-9336.002.patch
>
>
> The JobHistoryServer is leaking CLOSE_WAIT connections to DataNodes whenever 
> viewing a huge log file in JobhistoryServer. This happens only when the below 
> is configured.
> {code:java}
> yarn.log-aggregation.file-formats=IndexedFormat
> yarn.log-aggregation.file-controller.IndexedFormat.class=org.apache.hadoop.yarn.logaggregation.filecontroller.ifile.LogAggregationIndexedFileController
> yarn.log.server.url=http://jobhistory-host:19888/jobhistory/logs
> {code}
> On investigation, I found that the FSDataInputStream is not closed in 
> org.apache.hadoop.yarn.logaggregation.filecontroller.ifile.IndexedFileAggregatedLogsBlock
>  . Since this block is called every time the Jobhistory page displays the 
> logs, CLOSE_WAIT connections to DataNodes keep on increasing in 
> JobHistoryServer.
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to