[
https://issues.apache.org/jira/browse/CASSANDRA-6541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13945527#comment-13945527
]
Benedict commented on CASSANDRA-6541:
-------------------------------------
bq. Yes, so the statement "this HotSpot regression is not-worked-around in all
extant versions of Cassandra since the dawn of time" is correct. Thanks for the
clarification!
No, it isn't, since the hot spot bug did not exist since the dawn of time. As
such the from-version is ill-defined, but probably the most sensible definition
requires determining the from-version-for-hotspot, which we have yet to manage,
and aligning that with C* releases.
> New versions of Hotspot create new Class objects on every JMX connection
> causing the heap to fill up with them if CMSClassUnloadingEnabled isn't set.
> -----------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-6541
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6541
> Project: Cassandra
> Issue Type: Bug
> Components: Config
> Reporter: jonathan lacefield
> Assignee: Brandon Williams
> Priority: Minor
> Fix For: 1.2.16, 2.0.6, 2.1 beta2
>
>
> Newer versions of Oracle's Hotspot JVM , post 6u43 (maybe earlier) and 7u25
> (maybe earlier), are experiencing issues with GC and JMX where heap slowly
> fills up overtime until OOM or a full GC event occurs, specifically when CMS
> is leveraged. Adding:
> {noformat}
> JVM_OPTS="$JVM_OPTS -XX:+CMSClassUnloadingEnabled"
> {noformat}
> The the options in cassandra-env.sh alleviates the problem.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
