[
https://issues.apache.org/jira/browse/HDFS-6694?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14100903#comment-14100903
]
Arpit Agarwal commented on HDFS-6694:
-------------------------------------
So this indicates that the build slaves should be configured with a numfiles
limit of 60000. Is that correct?
https://github.com/apache/toolchain/blob/master/roles/common/files/limits.conf
That doesn't match with the 1024 limit that Yongjun noted earlier. [~gkesavan],
do you know where the numfiles limits for the build slaves is configured?
Thanks.
> TestPipelinesFailover.testPipelineRecoveryStress tests fail intermittently
> with various symptoms
> ------------------------------------------------------------------------------------------------
>
> Key: HDFS-6694
> URL: https://issues.apache.org/jira/browse/HDFS-6694
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 3.0.0
> Reporter: Yongjun Zhang
> Assignee: Yongjun Zhang
> Priority: Blocker
> Fix For: 2.6.0
>
> Attachments: HDFS-6694.001.dbg.patch,
> org.apache.hadoop.hdfs.server.namenode.ha.TestPipelinesFailover-output.txt,
> org.apache.hadoop.hdfs.server.namenode.ha.TestPipelinesFailover.txt
>
>
> TestPipelinesFailover.testPipelineRecoveryStress tests fail intermittently
> with various symptoms. Typical failures are described in first comment.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
