Mailing-List: contact dev-help@activemq.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@activemq.apache.org
Date: Fri, 5 Jul 2013 13:09:50 +0000 (UTC)
From: "Timothy Bish (JIRA)" <jira@apache.org>
To: dev@activemq.apache.org
Message-ID: <JIRA.12656338.1373012684476.3862.1373029790474@arcas>
In-Reply-To: <JIRA.12656338.1373012684476@arcas>
References: <JIRA.12656338.1373012684476@arcas>
Subject: [jira] [Commented] (AMQ-4618) Failover configuration establishes
 more and more connections to inactive ActiveMQ instance
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/AMQ-4618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13700650#comment-13700650 ] 

Timothy Bish commented on AMQ-4618:
-----------------------------------

I'd recommend you test against the latest release, or the latest 5.9.SNAPSHOT build to see if you can reproduce this.  
                
> Failover configuration establishes more and more connections to inactive ActiveMQ instance
> ------------------------------------------------------------------------------------------
>
>                 Key: AMQ-4618
>                 URL: https://issues.apache.org/jira/browse/AMQ-4618
>             Project: ActiveMQ
>          Issue Type: Bug
>    Affects Versions: 5.6.0
>            Reporter: Roberto Mier
>
> We have a big problem in a production environment usign ActiveMQ 5.6.0:
> We have two ActiveMQ instances started up and a consumer application reading from them. The configuration is set to failover so that consumer application reads JMS messages from first instance and changes to second instance only if first one is not reachable.
> After some days, we have seen that there is only one connection established to first ActiveMQ instance, but there are more than 80 established connections to second instance (the backup one). Every day more and more connections are created to backup instance until "java.net.SocketException: Too many open files" exception is thrown in consumer application because the number of file descriptors is huge
> Between ActiveMQ instances and consumer application there is a F5 with a timeout of 300 seconds. Could it be responsible of this bad behaviour?. I think that when F5 detects inactivity in a connection for 300 seconds, it will close it, but ActiveMQ will open a new one, so there should be only one connection at the same time, not growing the number of them, isn't it?
> The failover configuration in client is set to:
> jms.broker.url=failover:(tcp://activeMQHost1:61616,tcp://activeMQHost2:61616)?randomize=false&timeout=3000&priorityBackup=true
> Is there any additional parameter to set to solve this problem?
> Thanks in advance

--
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