[
https://issues.apache.org/jira/browse/HDFS-6195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13961984#comment-13961984
]
Mit Desai commented on HDFS-6195:
---------------------------------
While cleaning up the containers,
{code}
while (cleanedSize < allocatedSize && waitCount++ < 200) {
Thread.sleep(100);
resp = nm.nodeHeartbeat(true);
cleaned = resp.getContainersToCleanup();
cleanedSize += cleaned.size();
}
{code}
The test sometimes cannot do the complete cleanup and some of the 10000
containers cannot be cleaned up. Resulting an assertion error at
{{Assert.assertEquals(allocatedSize, cleanedSize);}}.
This test has been failing in our nightly builds since couple of days. I was
able to reproduce this consistently on eclipse but not using maven. I think
this is an environment issue so cannot be reproduced everywhere.
As a fix, I have increased the thread sleep time in the while loop. Which will
give some extra time for the container cleanup. And as there is also a check in
the while loop for the allocated size and cleaned size, the test will not
always take up all cycles in the loop.
> TestRMApplicationHistoryWriter#testRMWritingMassiveHistory runs slow and
> intermittently fails on trunk and branch2
> ------------------------------------------------------------------------------------------------------------------
>
> Key: HDFS-6195
> URL: https://issues.apache.org/jira/browse/HDFS-6195
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 3.0.0, 2.5.0
> Reporter: Mit Desai
> Assignee: Mit Desai
> Fix For: 3.0.0, 2.5.0
>
> Attachments: HDFS-6195.patch
>
>
> The test has 10000 containers that it tries to cleanup.
> The cleanup has a timeout of 20000ms in which the test sometimes cannot do
> the cleanup completely and gives out an Assertion Failure.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
