[
https://issues.apache.org/jira/browse/SOLR-12807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16631479#comment-16631479
]
Mine Orange commented on SOLR-12807:
------------------------------------
Well,as the SOLR-10420 shows,the childwatcher leak in DistributedQueue may
eventually lead to OOM of my instance,i have tested the 6.6 version and
compared it with the 6.1 version,focusing on the registration of the
childwatcher. There is no repeated registration problem of watcher in 6.6
version.
> out of memory error due to a lot of zk watchers in solr cloud
> --------------------------------------------------------------
>
> Key: SOLR-12807
> URL: https://issues.apache.org/jira/browse/SOLR-12807
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: SolrCloud
> Affects Versions: 6.1
> Reporter: Mine Orange
> Priority: Major
>
> Analyzing the dump file,we found a lot of watchers in childWatches of
> ZKWatchManager,nearly 1.8G,the znode of childWatches is
> /overseer/collection-queue-work,confirm that it is not because of the
> frequent use of collection API and the network is normal.
> The instance is the overseer leader of a solr cluster and did not restart for
> more than a year,suspect that the watchers grow with time.
> Our solr version is 6.1 and zookeeper version is 3.4.9.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
