You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@activemq.apache.org by AndyG <an...@orprovision.com> on 2012/11/28 11:02:48 UTC
Re: InactivityMonitor - Creating too frequent threads
I agree. I am using ActiveMQ within OpenEJB and under our normal load I have
about 140 threads alive. Just about the only thread deaths I have are the
'ActiveMQ InactivityMonitor Worker' and 'ActiveMQ BrokerService[name]
Task-[count].
I tend to wait for a stable snapshot and then change the pool value to 2
minutes in the AbstractInactivityMonitor.java (default 10 seconds) and
TaskRunnerFactory.java (default 30 seconds) and so far I have not seen the
thread count creep up due to the fact that the pools are unbounded.
It would make life easier if these values were a configuration option or at
least taken from a System property.
Maybe a JIRA is in order?
Andy.
--
View this message in context: http://activemq.2283324.n4.nabble.com/InactivityMonitor-Creating-too-frequent-threads-tp4656752p4659873.html
Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
Re: InactivityMonitor - Creating too frequent threads
Posted by AndyG <an...@orprovision.com>.
Here we go:
https://issues.apache.org/jira/browse/AMQ-4205
Sorry, was away from the box for a while. And thanks for the OpenEJB/TomEE
link on the site.
Andy.
--
View this message in context: http://activemq.2283324.n4.nabble.com/InactivityMonitor-Creating-too-frequent-threads-tp4656752p4660151.html
Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
Re: InactivityMonitor - Creating too frequent threads
Posted by AndyG <an...@orprovision.com>.
Also, just to be clear, this is not a task timeout. It is the pool
terminating 'idle' threads due to 'keepAliveTime' expiration.
--
View this message in context: http://activemq.2283324.n4.nabble.com/InactivityMonitor-Creating-too-frequent-threads-tp4656752p4660153.html
Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
Re: InactivityMonitor - Creating too frequent threads
Posted by Claus Ibsen <cl...@gmail.com>.
Hi Andy
Yeah sure create a JIRA ticket.
When I looked into this initially I didn't want to "change" default
and cause any side-effects.
As I am still new in some areas of ActiveMQ source code.
I think I have seen others have a + 1.5 x as their thread timeout
default value for scheduled tasks.
So if the task runs for 30 sec, then it has a timeout of + 45 sec = 75 sec.
Allowing people to configure this would at least allow people to go
back to the old behavior
in case for som "odd" reason that worked better.
Another reason why I would like to avoid this thread creation /
termination all the time.
Is that I have seen platforms crash the JVM due OOME when
allocating/creating the thread.
Anyway a JIRA and pactches is of course as usually always welcome at
ASF projects.
On Wed, Nov 28, 2012 at 11:02 AM, AndyG <an...@orprovision.com> wrote:
> I agree. I am using ActiveMQ within OpenEJB and under our normal load I have
> about 140 threads alive. Just about the only thread deaths I have are the
> 'ActiveMQ InactivityMonitor Worker' and 'ActiveMQ BrokerService[name]
> Task-[count].
>
> I tend to wait for a stable snapshot and then change the pool value to 2
> minutes in the AbstractInactivityMonitor.java (default 10 seconds) and
> TaskRunnerFactory.java (default 30 seconds) and so far I have not seen the
> thread count creep up due to the fact that the pools are unbounded.
>
> It would make life easier if these values were a configuration option or at
> least taken from a System property.
>
> Maybe a JIRA is in order?
>
> Andy.
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.nabble.com/InactivityMonitor-Creating-too-frequent-threads-tp4656752p4659873.html
> Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
--
Claus Ibsen
-----------------
Red Hat, Inc.
FuseSource is now part of Red Hat
Email: cibsen@redhat.com
Web: http://fusesource.com
Twitter: davsclaus
Blog: http://davsclaus.com
Author of Camel in Action: http://www.manning.com/ibsen