[jboss-jira] [JBoss JIRA] (JGRP-2068) Problems with JBOSS cluster

[Fouad Zoghlami (JIRA)]

    [ https://issues.jboss.org/browse/JGRP-2068?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13241234#comment-13241234 ] 

Fouad Zoghlami commented on JGRP-2068:
--------------------------------------

could this be the same problem as:

the problem seems to be a problem with the jgroups.
Out of the log files:
...
2015-12-09 16:56:11,132 FATAL CloserThread [org.jgroups.JChannel] local_addr is null; cannot connect
2015-12-09 16:56:11,132 ERROR CloserThread [org.jgroups.JChannel] failure reconnecting to channel, retrying
org.jgroups.ChannelException: local_addr is null
	at org.jgroups.JChannel.startStack(JChannel.java:1631)
	at org.jgroups.JChannel.connect(JChannel.java:366)
	at org.jgroups.JChannel$CloserThread.run(JChannel.java:2046)
...

Jgroups is not able to reconnect.

I found the following articles:


https://developer.jboss.org/message/8089

https://issues.jboss.org/browse/JGRP-1006


According to the second article it's fixed with jgroups 2.6.11


> Problems with JBOSS cluster
> ---------------------------
>
>                 Key: JGRP-2068
>                 URL: https://issues.jboss.org/browse/JGRP-2068
>             Project: JGroups
>          Issue Type: Bug
>    Affects Versions: 2.6.10
>            Reporter: Fouad Zoghlami
>            Assignee: Bela Ban
>         Attachments: the master.zip, the slave.zip
>
>
> We are using a JBOSS cluster for the Process Server component. One node (the master) is used for Process Workplace and a second node is used by an external application to start new process instances from the scanner input.
> Recently they encountered a problem where one of the 2 nodes started to give problems. When we look in the logs it looks like there is some kind of communication problem between the 2 nodes in the cluster. From that point it looks like the cluster is repaired, but from that point the process server on one of the two nodes begins to give problems when users try to send on workitems. It looks like the locking state is out of sync or something like that. In all situation we have been able to fix the issue by restarting jboss on both nodes and clearing the /tmp, /data and /work folders
> Please see the logs file.
> Issue happened around 2015-11-12 10:48:35. It happened also 2 weeks ago and a couple of months ago.
> Please assist in finding the cause of this behavior and how we can prevent this in future.



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)