[
https://issues.apache.org/jira/browse/HDFS-14863?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16938012#comment-16938012
]
David Mollitor commented on HDFS-14863:
---------------------------------------
Different unit tests failed on the second Yetus run. Flaky tests.
This particular data structure is accessed in a few places, but this is the
only place it is synchronized on. I just don't see a reason for it and it's
not documented anywhere as to why this may be the case.
> Remove Synchronization From BlockPlacementPolicyDefault
> -------------------------------------------------------
>
> Key: HDFS-14863
> URL: https://issues.apache.org/jira/browse/HDFS-14863
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: block placement
> Affects Versions: 3.2.0
> Reporter: David Mollitor
> Assignee: David Mollitor
> Priority: Major
> Attachments: HDFS-14863.1.patch, HDFS-14863.2.patch
>
>
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockPlacementPolicyDefault.java#L1010
> The {{clusterMap}} has its own internal synchronization. Also, these are
> only read operations so any changes applied to the {{clusterMap}} from
> another thread will be applied since no other thread synchronizes on the
> {{clusterMap}} itself (that I could find).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
