You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by CurtisRopp <cr...@teamropp.com> on 2011/01/26 20:13:07 UTC
ActiveMQ 5.1.0 queue SawTooth pattern
I am run ActiveMQ 5.1.0 moving roughly 120 messages per second into one queue
and then filtering and copying to 4 unique queues using camel where similar
consumers are consuming those messages.
Over a period of time the memory trend starts to grow and we have learned
that around a certain point we need to go in and restart amq (maybe a
different conversation) when we do this restart and about 30% of the time
one of the unique queues will go into a sawtooth pattern where messages
don't get delivered until the peak we have seen this between 20k and 240k at
the top and then they get consumed.
Have others ran into this issue, is it a known defect that I can't find, or
the bigger question does anyone know how to fix?
Thanks,
Curtis
--
View this message in context: http://activemq.2283324.n4.nabble.com/ActiveMQ-5-1-0-queue-SawTooth-pattern-tp3238641p3238641.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: ActiveMQ 5.1.0 queue SawTooth pattern
Posted by CurtisRopp <cr...@teamropp.com>.
Dejan,
Thanks for you reply yes we have spent time testing 5.3.x, 5.4, and now
5.4.2 with 5.4.2 being what we believe has been the more stable release
since 5.1.0. So far this appears to only happen in production and I don't
have a repeatable test case for this pattern, wish I did. Does that mean
that similar issues have not been seen before?
Regards,
Curtis
--
View this message in context: http://activemq.2283324.n4.nabble.com/ActiveMQ-5-1-0-queue-SawTooth-pattern-tp3238641p3242937.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: ActiveMQ 5.1.0 queue SawTooth pattern
Posted by Dejan Bosanac <de...@nighttale.net>.
Hi Curtis,
did you test some of the later versions? It be good if you could test a
current snapshot and produce a test case if the problem still exists.
Regards
--
Dejan Bosanac
-----------------
FuseSource - The experts in open source integration and messaging.
Email: dejanb@fusesource.com
Web: http://fusesource.com
Twitter: http://twitter.com/dejanb
ActiveMQ in Action - http://www.manning.com/snyder/
Blog - http://www.nighttale.net
On Wed, Jan 26, 2011 at 8:13 PM, CurtisRopp <cr...@teamropp.com> wrote:
>
> I am run ActiveMQ 5.1.0 moving roughly 120 messages per second into one
> queue
> and then filtering and copying to 4 unique queues using camel where similar
> consumers are consuming those messages.
>
> Over a period of time the memory trend starts to grow and we have learned
> that around a certain point we need to go in and restart amq (maybe a
> different conversation) when we do this restart and about 30% of the time
> one of the unique queues will go into a sawtooth pattern where messages
> don't get delivered until the peak we have seen this between 20k and 240k
> at
> the top and then they get consumed.
>
> Have others ran into this issue, is it a known defect that I can't find, or
> the bigger question does anyone know how to fix?
>
> Thanks,
>
> Curtis
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-5-1-0-queue-SawTooth-pattern-tp3238641p3238641.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>