activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Duro <duroku...@zoznam.sk>
Subject Re: Memory limit reached on persistent queue
Date Wed, 09 Dec 2009 15:09:59 GMT

Thanks for answer but unfortunatelly none of these settings work for me
despite that message size is only cca 25kb.


Thierry H wrote:
> 
> I 
> I had the same problem , and it seems that you have 2 issues on the 5.3 on
> the subject.
> I think your problem come from extensive memory usage related to these
> issues.
> I've workarounded the problem by changing the DestinationPolicy of the
> queue 
> deactivate MessageExpiration scheduled scan : expireMessagesPeriod="0"
> reduce the prefetch and maxPageSize policies which were to high for me (As
> my messages are big)
> maxPageSize="50" queuePrefetch="300"
> 
> I hope that help 
> 
> Regards
>  
> Hi,
> I'am using ActiveMQ 5.3 with kahaPersistenceAdapter my test case is:
> - slow consumer(s) 
> - fast producer
> - persistent queue
> - producerFlowControl="true"
> - Win XP (Sles10)
> 
> after couple thousand of massages stored massages i can find following
> massge on console:
> 
> INFO | Usage Manager memory limit reached on queue://slowConsumerDest.
> Producers will be throttled to the rate at which messages are removed from
> this destination to prevent flooding it. See
> http://activemq.apache.org/producer-flow-control.html for more info
> 
> -in jconsole i see that CursorPercentUsage reached value 100.
> from this point no more messages are persisted to disc and producer si
> sending only as fast as consumer(s) are receiving. 
> 
> My question is why no more messages are stored when i'm using persistent
> messaging? 
> 
> I've try same configuration on ActiveMQ 5.3 snapshot (rev. number 747951)
> - CursorPercentUsage reached 70 and incomming messages were stored as i've
> expected. 
> This mechanism works with Amq 5.3 in case amqPersistenceAdapter is used,
> for kahaPersistenceAdapter when topic with durable subscriber used instead
> of queue and for queue in case it has no consumers.
> Using of kahadb is not an option in my case while handling of persistent
> massages is several times slower.
> 
> Regards,
> Juraj
> 
>  
> 
> 



-- 
View this message in context: http://old.nabble.com/Memory-limit-reached-on-persistent-queue-tp26709801p26711767.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.


Mime
View raw message