You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Erick Erickson (JIRA)" <ji...@apache.org> on 2016/02/09 15:53:18 UTC
[jira] [Resolved] (SOLR-8663) Multiple searchers open for a single
core causing memory issues
[ https://issues.apache.org/jira/browse/SOLR-8663?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Erick Erickson resolved SOLR-8663.
----------------------------------
Resolution: Invalid
First, please discuss this kind of thing on the user's list and/or the dev list before raising a JIRA, we try to keep JIRAs for code changes. You'll get more responses faster there.
This is almost certainly a problem in your custom code where you're not properly closing searchers. Indeed, when you keep a searcher open it chews up memory.
As for the warmup, warming a bunch of searchers at once usually means you're opening them too rapidly and is an anti-pattern.
Closing this until we have an indication that this is a Solr issue rather than a custom code issue.
> Multiple searchers open for a single core causing memory issues
> ---------------------------------------------------------------
>
> Key: SOLR-8663
> URL: https://issues.apache.org/jira/browse/SOLR-8663
> Project: Solr
> Issue Type: Bug
> Affects Versions: 4.6
> Reporter: Dhritiman Das
>
> For a single core, we are having multiple searchers open. Although the current searcher points to the latest searcher opened, some old searchers are not getting cleaned up , they are holding their cache and ultimately causing memory issues.
> Possibly for cores which are taking a lot of time to warmup their cache the replication interval is overlapping with warmup time taken for the last searcher open during last replication.
> However any idea why sometimes old searchers are not getting freed ?
> Ex. Searchers from couple of days back are still seen in Solr console.
> ip:port/solr/admin.html#/<core-name>/plugins/core
> Is there any known bug around this for Solr 4.6 ?
> p.s.: we have custom code extending some Solr components at many places. What I am trying to understand also is what is the best place to start debugging this ?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org