[
https://issues.apache.org/jira/browse/HDFS-13675?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16512738#comment-16512738
]
Íñigo Goiri commented on HDFS-13675:
------------------------------------
All the tests passed
[here|https://builds.apache.org/job/PreCommit-HDFS-Build/24435/testReport/org.apache.hadoop.hdfs.tools/TestDFSAdminWithHA/]
and we went all the way down to 4 minutes and 27 seconds.
+1 on [^HDFS-13675.000.patch].
Committing after HDFS-13563 (we may need to rebase).
> Speed up TestDFSAdminWithHA
> ---------------------------
>
> Key: HDFS-13675
> URL: https://issues.apache.org/jira/browse/HDFS-13675
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs, namenode
> Reporter: Lukas Majercak
> Assignee: Lukas Majercak
> Priority: Major
> Attachments: HDFS-13675.000.patch
>
>
> Currently, TestDFSAdminWithHA takes about 10 minutes to finish. The main
> culprits are two tests:
> testListOpenFilesNN1DownNN2Down
> testSetBalancerBandwidthNN1DownNN2Down
>
> that each take 3~ minutes to finish. This is because they both expect to fail
> to connect to 2 namenodes, but the client retry policy has way too many
> retries and exponential backoffs.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
