[
https://issues.apache.org/jira/browse/YARN-1630?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13881507#comment-13881507
]
Sandy Ryza commented on YARN-1630:
----------------------------------
Can the config be added as a timeout instead of a number of polls?
Also, a couple nits:
Timeout should default to -1, meaning forever
No need for message var. just LOG.info directly.
An (Yarn)exception should be thrown to indicate that the operation didn't
complete in time. Otherwise clients might think it had completed succesfuly.
> Unbounded waiting for response in YarnClientImpl.java causes thread to hang
> forever
> -----------------------------------------------------------------------------------
>
> Key: YARN-1630
> URL: https://issues.apache.org/jira/browse/YARN-1630
> Project: Hadoop YARN
> Issue Type: Bug
> Components: client
> Affects Versions: 2.2.0
> Reporter: Aditya Acharya
> Assignee: Aditya Acharya
> Attachments: diff.txt
>
>
> I ran an MR2 application that would have been long running, and killed it
> programmatically using a YarnClient. The app was killed, but the client hung
> forever. The message that I saw, which spammed the logs, was "Watiting for
> application application_1389036507624_0018 to be killed."
> The RM log indicated that the app had indeed transitioned from RUNNING to
> KILLED, but for some reason future responses to the RPC to kill the
> application did not indicate that the app had been terminated.
> I tracked this down to YarnClientImpl.java, and though I was unable to
> reproduce the bug, I wrote a patch to introduce a bound on the number of
> times that YarnClientImpl retries the RPC before giving up.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
