[jira] Commented: (AMQ-1069) Setting maxInactivityDuration on broker to broker connections to value greater than 30 seconds doesn't work

["james strachan (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/activemq/browse/AMQ-1069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_38680 ] 

james strachan commented on AMQ-1069:
-------------------------------------

We've done a bunch of changes in 4.1.1 in this area to fix a few different related issues. I wonder could you see if the current 4.1.1 build fixes your issue?

http://people.apache.org/repo/m2-snapshot-repository/org/apache/activemq/apache-activemq/4.1-SNAPSHOT/

> Setting maxInactivityDuration on broker to broker connections to value greater than 30 seconds doesn't work
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: AMQ-1069
>                 URL: https://issues.apache.org/activemq/browse/AMQ-1069
>             Project: ActiveMQ
>          Issue Type: Bug
>    Affects Versions: 4.1.0
>         Environment: Linux RHEL 4 with Java 1.5
>            Reporter: Eric Wood
>            Priority: Critical
>
> We are running AMQ in a configuration with embedded brokers on both the producer and consumer (broker to broker network communication).  We have an extremely high latency network, which results in a large number of inactivity timeouts.  Once we discovered how to set the timeout, we tried to set it to a value of 100 seconds which we believe will work better with our configuration and network.  However, we observed that timeouts were still occuring after 30 seconds of inactivity.  We tried setting the inactivity timeout value on both brokers and got the same result.  We tried setting it to a value less than 30 (i.e. 10)  and we saw timeouts occuring on that interval.  Because of this, we believe that there must be a limit on the inactivity timeout value.  If there is such a limit, could it be increased.? The 30 second value causes a large number of timeouts, which we believe causes a lower effective throughput than is truly possible with JMS/AMQ.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.