[jboss-jira] [JBoss JIRA] Commented: (JGRP-1283) reproducing outofmemoryerror

[Bela Ban (JIRA)]

    [ https://issues.jboss.org/browse/JGRP-1283?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12580838#comment-12580838 ] 

Bela Ban commented on JGRP-1283:
--------------------------------

Ronald, we should discuss questions of conceptual nature on the dev or users mailing list, as others will benefit, too...

Yes, group communication is hampered by unresponsive members. For example: stability (STABLE) will accumulate messages, as its not able to reach consensus among its members because there's no response from the GC'ed member. Ditto for flow control.

10 seconds of GC is not an issue. However, you should strive for more, and smaller, GCs. In general, failure detection limits the amount of time an unresponsive member can hamper the rest of the cluster. After the timeout elapses, an unresponsive member is kicked out of the cluster and communication continues.

> reproducing outofmemoryerror
> ----------------------------
>
>                 Key: JGRP-1283
>                 URL: https://issues.jboss.org/browse/JGRP-1283
>             Project: JGroups
>          Issue Type: Bug
>            Reporter: ronald yang
>            Assignee: Bela Ban
>            Priority: Minor
>             Fix For: 2.12
>
>         Attachments: Leader.java, oome.tgz, tcp-nio.xml
>
>


--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira