[
https://issues.apache.org/jira/browse/CASSANDRA-3991?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13738656#comment-13738656
]
Arya Goudarzi commented on CASSANDRA-3991:
------------------------------------------
Back in November 2012, we had a 12 node cluster of m1.xlarge EC2 instances. We
had about 30 ColumnFamilies on the cluster. Each node had 8Gb heap and we
didn't change the defaults. We had a workload of about 3K writes and 1.5k reads
per second on the cluster as a result of running some batch jobs. The heap just
got filled up all the way to the point that GC was not happening any more. At
this time gossip was still running and announcing but all operations to the
node were timing out. Top showed JSVC is using 60Gb of virtual memory, and I
supposed that was mmaped files. I suppose if you stress C* on a debian node you
can reproduce this easily, although I never got a chance to try it
independently. We use Ubuntu 10.4 LTS. At the time this was happening we had C*
1.1.x. Now we have 1.2.5. I ended up replacing JSVC with start-stop-daemon.
> Investigate importance of jsvc in debian packages
> -------------------------------------------------
>
> Key: CASSANDRA-3991
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3991
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Brandon Williams
> Assignee: Eric Evans
> Priority: Minor
> Fix For: 2.1
>
>
> jsvc seems to be buggy at best. For instance, if you set a small heap like
> 128M it seems to completely ignore this and use as much memory as it wants.
> I don't know what this is buying us over launching /usr/bin/cassandra
> directly like the redhat scripts do, but I've seen multiple complaints about
> its memory usage.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
