You are viewing a plain text version of this content. The canonical link for it is here.

Posted to users@activemq.apache.org by "J. Matthew Pryor" <jm...@observant.com.au> on 2010/01/03 03:07:56 UTC

Massive accumulation of ConcurrentHashMap in heap

Hi,

I am trying to track down the reason for a massive accumulation of  
ConcurrentHashMap in the heap of our long running Java process.

We are running Java 1.6 so I am pretty sure the known issue relating  
to the 1.5 JVM doesn't apply here.

The URL below is a link to a screen shot of the heap analysis, showing  
the accumulated memory. It seems to be root in a series or region  
brokers?

http://img160.yfrog.com/img160/1049/heap.png

I can also see that some JMX classes are in the root, could the reason  
for the accumulation be related to having JMX enabled?

As far as we can tell, the queues are being serviced normally, and  
this application generally runs for long periods of time without any  
noticeable memory leak.

I'd be greatly appreciative of any ideas for strategies to get more  
information about the source of the memory accumulation.

We're using Java 1.6 7 Active MQ 5.3

Thanks,
Matthew

Re: Massive accumulation of ConcurrentHashMap in heap

Posted by "J. Matthew Pryor" <jm...@observant.com.au>.

Thanks for the reply Rob,

I am not sure how we are accumulating these connections, could be  
anything to do with having turned JMX on?

We sometime have clients with intermittent TCP/IP network connection?  
Perhaps we have timeout issues or some other reason for retaining the  
connections?

Can you suggest any mods or config changes we can perform to help  
diagnosis?

We still have the dumped heap file, are there any other queries we can  
perform that might shed more light?

Thanks again

Matthew

On 03/01/2010, at 9:13 PM, Rob Davies wrote:

>
> On 3 Jan 2010, at 02:07, J. Matthew Pryor wrote:
>
>> Hi,
>>
>> I am trying to track down the reason for a massive accumulation of  
>> ConcurrentHashMap in the heap of our long running Java process.
>>
>> We are running Java 1.6 so I am pretty sure the known issue  
>> relating to the 1.5 JVM doesn't apply here.
>>
>> The URL below is a link to a screen shot of the heap analysis,  
>> showing the accumulated memory. It seems to be root in a series or  
>> region brokers?
>>
>> http://img160.yfrog.com/img160/1049/heap.png
>>
>> I can also see that some JMX classes are in the root, could the  
>> reason for the accumulation be related to having JMX enabled?
>>
>> As far as we can tell, the queues are being serviced normally, and  
>> this application generally runs for long periods of time without  
>> any noticeable memory leak.
>>
>> I'd be greatly appreciative of any ideas for strategies to get more  
>> information about the source of the memory accumulation.
>>
>> We're using Java 1.6 7 Active MQ 5.3
>>
>> Thanks,
>> Matthew
>>
> Yuk - I thought I had purged all the use of ConcurrentHashMap - it  
> looks like you have a lot of active connections ??
>
> Rob Davies
> http://twitter.com/rajdavies
> I work here: http://fusesource.com
> My Blog: http://rajdavies.blogspot.com/
> I'm writing this: http://www.manning.com/snyder/
>
>
>
>
>

Re: Massive accumulation of ConcurrentHashMap in heap

Posted by Rob Davies <ra...@gmail.com>.

On 3 Jan 2010, at 02:07, J. Matthew Pryor wrote:

> Hi,
>
> I am trying to track down the reason for a massive accumulation of  
> ConcurrentHashMap in the heap of our long running Java process.
>
> We are running Java 1.6 so I am pretty sure the known issue relating  
> to the 1.5 JVM doesn't apply here.
>
> The URL below is a link to a screen shot of the heap analysis,  
> showing the accumulated memory. It seems to be root in a series or  
> region brokers?
>
> http://img160.yfrog.com/img160/1049/heap.png
>
> I can also see that some JMX classes are in the root, could the  
> reason for the accumulation be related to having JMX enabled?
>
> As far as we can tell, the queues are being serviced normally, and  
> this application generally runs for long periods of time without any  
> noticeable memory leak.
>
> I'd be greatly appreciative of any ideas for strategies to get more  
> information about the source of the memory accumulation.
>
> We're using Java 1.6 7 Active MQ 5.3
>
> Thanks,
> Matthew
>
Yuk - I thought I had purged all the use of ConcurrentHashMap - it  
looks like you have a lot of active connections ??

Rob Davies
http://twitter.com/rajdavies
I work here: http://fusesource.com
My Blog: http://rajdavies.blogspot.com/
I'm writing this: http://www.manning.com/snyder/