[
https://issues.apache.org/jira/browse/YARN-5927?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15702603#comment-15702603
]
Miklos Szegedi commented on YARN-5927:
--------------------------------------
Thank you for the patch [~lewuathe]! I also suggest reducing
{code}Thread.sleep(2000);{code} to 10 milliseconds. That is the usual range of
these container events to occur, I think.
> BaseContainerManagerTest::waitForNMContainerState timeout accounting is not
> accurate
> ------------------------------------------------------------------------------------
>
> Key: YARN-5927
> URL: https://issues.apache.org/jira/browse/YARN-5927
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Miklos Szegedi
> Assignee: Kai Sasaki
> Priority: Trivial
> Labels: newbie
> Attachments: YARN-5917.01.patch
>
>
> See below that timeoutSecs is increased twice. We also do a sleep right away
> before even checking the observed value.
> {code}
> do {
> Thread.sleep(2000);
> ...
> timeoutSecs += 2;
> } while (!finalStates.contains(currentState)
> && timeoutSecs++ < timeOutMax);
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
