[
https://issues.apache.org/jira/browse/HDFS-5589?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13861782#comment-13861782
]
Colin Patrick McCabe commented on HDFS-5589:
--------------------------------------------
Good work-- it's great to have finally found that bug.
{code}
+ DatanodeDescriptor datanode = chooseDatanodeForCaching(possibilities,
+ blockManager.getDatanodeManager().getStaleInterval());
+ assert (datanode != null);
{code}
Don't we need to handle the case where {{datanode == null}} here? It seems
like that will come up if, for example, we have no space available on any
datanode. It might be better to move the code from {{if
(possibilities.isEmpty()) }} if statement into a new check for
{{chooseDataNodeForCaching == null}}. It would be nice to see for this case
too-- just fill up the DN caches, or set them to 0...
> Namenode loops caching and uncaching when data should be uncached
> -----------------------------------------------------------------
>
> Key: HDFS-5589
> URL: https://issues.apache.org/jira/browse/HDFS-5589
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: caching, namenode
> Affects Versions: 3.0.0
> Reporter: Andrew Wang
> Assignee: Andrew Wang
> Attachments: hdfs-5589-1.patch
>
>
> This was reported by [~cnauroth] and [~brandonli], and [~schu] repro'd it too.
> If you add a new caching directive then remove it, the Namenode will
> sometimes get stuck in a loop where it sends DNA_CACHE and then DNA_UNCACHE
> repeatedly to the datanodes where the data was previously cached.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
