[ 
https://issues.apache.org/jira/browse/AMQ-3710?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17270005#comment-17270005
 ] 

Dee Moore commented on AMQ-3710:
--------------------------------

We hit the same issue on ActiveMQ 5.14.2. It was configured to delete inactive 
destinations.

We solved this by reviewing our topics - we were making transient topics such 
as /topic/topic.prefix.<uuid>.entry, which sent small amounts of data and then 
became inactive. However, one of our subscribers was subscribing to 
/topic/topic.prefix.>, and this was enough to keep every topic active, and 
resulted in steady growth of db.data, with only one active .log file.

Even though we found a solution, it would be very nice to be able to mark a 
channel as "This channel is inactive when messages have not been sent for x 
amount of time and should be deleted, whether or not there are active 
subscribers". I cannot find a way to do this - having a subscriber, even on a 
wild card basis, is enough to keep a topic active.

> Activemq broker hangs
> ---------------------
>
>                 Key: AMQ-3710
>                 URL: https://issues.apache.org/jira/browse/AMQ-3710
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker
>    Affects Versions: 5.4.2
>         Environment: cent os linux version 5, 64 bit os, java 1.6 update 21
>            Reporter: Krishnadasan T S
>            Priority: Major
>
> I am using activeMQ 5.4.2 as separate process in my production. more than 
> 2000 topics and 600 clients are connecting to broker. Heap size (Old 
> generation) area is taking nearly 1 gb memory. Total memory allocated to jvm 
> is 2.5 gb.
>   Activemq is getting hanged 2 or 3 time in a day. by monitoring in JMX 
> console What we noticed is that at that instance of time nearly 100 new 
> threads are created and all these threads are NIO related threads. Once It 
> happened at the time of GC(we gave a perform gc from jmx for testing). It 
> also happened at normal time also.
> can you give any suggestion about what can be the issue.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to