Re: JMS publish consuming sockets?
On Sat, Jul 11, 2009 at 1:26 AM, huntc wrote: > > > Claus Ibsen-2 wrote: >> >> Are your routes InOnly? Eg you do not consume from a JMS queue? >> > > Yes I do have some in-only routes. Could the maxMessagesPerTask default > issue cause an unwieldily number of connections to be made with the broker? This option is part of Spring JMS and could also be used/affect when sending only. So I advice to configure all your JMS endpoints in 1.6.1 with maxMessagesPerTask=-1 > > Also is there any feel for when 1.6.2 is out or better still 2.0? It is > probably time I got off the 1.x source base given 2.0 is probably where most > focus is nowadays. 1.6.x should be good enough for production usage. Out intention is of course to support it as good as we can and do important bug fixes. But out own focus and energy has been on 2.0 for quite a while now. I recommend to try out 2.0m2 or better yet 2.0 trunk when you got the time. Maybe giving your existing application a test drive with it. The more feedback we can get with X apps not working / working after a upgrade the better. But we picked up a recent discussion on some API changes for 2.0 so it would take another 1-2 weeks before a RC can be cut, as opposed to I was hoping the RC could have been cut this weekend. But anyway its feature complete. But the API is something we gotta cease the moment and mold before its GA. > -- > View this message in context: > http://www.nabble.com/JMS-publish-consuming-sockets--tp24401653p24435619.html > Sent from the Camel - Users (activemq) mailing list archive at Nabble.com. > > -- Claus Ibsen Apache Camel Committer Open Source Integration: http://fusesource.com Blog: http://davsclaus.blogspot.com/ Twitter: http://twitter.com/davsclaus
Re: JMS publish consuming sockets?
Claus Ibsen-2 wrote: > > Are your routes InOnly? Eg you do not consume from a JMS queue? > Yes I do have some in-only routes. Could the maxMessagesPerTask default issue cause an unwieldily number of connections to be made with the broker? Also is there any feel for when 1.6.2 is out or better still 2.0? It is probably time I got off the 1.x source base given 2.0 is probably where most focus is nowadays. -- View this message in context: http://www.nabble.com/JMS-publish-consuming-sockets--tp24401653p24435619.html Sent from the Camel - Users (activemq) mailing list archive at Nabble.com.
Re: JMS publish consuming sockets?
Hi
Are your routes InOnly? Eg you do not consume from a JMS queue?
There was a bad default in camel-jms. See the warning on the JMS component page.
Maybe it can help.
And have you tried asking on the ActiveMQ forums. Must be an issue
they have heard before.
AMQ have a ton of settings to adjust.
On Thu, Jul 9, 2009 at 2:09 AM, huntc wrote:
>
> Hi guys,
>
> I've just noticed a problem with my 1.6.0 based application. I believe it
> may be establishing a connection with AMQ for the purposes of publishing on
> a topic, and then while retaining that connection, opening another
> connection the next time it needs to publish. This is just a hypothesis but
> after a good period of time (like a month or two), AMQ runs out of file
> descriptors and complains thus:
>
>
> Could not accept connection : Too many open files
>
>
> Here's my Camel route:
>
>
> from("direct:publishCAT21FeatureCollection").process(
> cat21FeatureCollectionToGMLTransformer).to(
>
> "activemq-sender:topic:com.classactionpl.javaFlightTopic");
>
>
>
> My spring config for AMQ is (activemq-receiver is defined but actually not
> used by my app at the moment):
>
>
> <bean id="activeMQConnectionFactory"
> class="org.apache.activemq.ActiveMQConnectionFactory">
> <property name="brokerURL"
> value="${com.classactionpl.asterixservices.brokerurl}"
> />
> </bean>
> <bean id="activemq-receiver"
> class="org.apache.activemq.camel.component.ActiveMQComponent">
> <property name="connectionFactory"
> ref="activeMQConnectionFactory"
> />
> </bean>
> <bean id="activemq-sender"
> class="org.apache.activemq.camel.component.ActiveMQComponent">
> <property name="connectionFactory">
> <bean
> class="org.apache.activemq.pool.PooledConnectionFactory">
> <constructor-arg
> ref="activeMQConnectionFactory" />
> </bean>
> </property>
> </bean>
>
>
> and the broker url:
>
>
> com.classactionpl.asterixservices.brokerurl=failover:tcp://localhost:61616
>
>
> Here's what I see in my Camel application's log file:
>
>
> 2009-07-09 09:47:09,607 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully reconnected to tcp://localhost:61616
> 2009-07-09 09:47:25,123 [127.0.0.1:61616] WARN FailoverTransport
> - Transport failed to tcp://localhost:61616 , attempting to automatically
> reconnect due to: java.net.SocketException: Connection reset
> 2009-07-09 09:47:25,132 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully reconnected to tcp://localhost:61616
> 2009-07-09 09:48:35,279 [127.0.0.1:61616] WARN FailoverTransport
> - Transport failed to tcp://localhost:61616 , attempting to automatically
> reconnect due to: java.io.EOFException
> 2009-07-09 09:48:35,291 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully reconnected to tcp://localhost:61616
> 2009-07-09 09:48:35,337 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully connected to tcp://localhost:61616
> 2009-07-09 09:49:35,478 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully connected to tcp://localhost:61616
> 2009-07-09 09:51:28,171 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully connected to tcp://localhost:61616
> 2009-07-09 09:52:55,163 [ActiveMQ Task ] INFO FailoverTransport
> - Successfully connected to tcp://localhost:61616
>
>
> I wouldn't expect so many failovers although it might just be the
> PooledConnectionFactory filling up its pool of connections.
>
> Any ideas?
>
> Kind regards,
> Christopher
> --
> View this message in context:
> http://www.nabble.com/JMS-publish-consuming-sockets--tp24401653p24401653.html
> Sent from the Camel - Users (activemq) mailing list archive at Nabble.com.
>
--
Claus Ibsen
Apache Camel Committer
Open Source Integration: http://fusesource.com
Blog: http://davsclaus.blogspot.com/
Twitter: http://twitter.com/davsclaus
JMS publish consuming sockets?
Hi guys,
I've just noticed a problem with my 1.6.0 based application. I believe it
may be establishing a connection with AMQ for the purposes of publishing on
a topic, and then while retaining that connection, opening another
connection the next time it needs to publish. This is just a hypothesis but
after a good period of time (like a month or two), AMQ runs out of file
descriptors and complains thus:
Could not accept connection : Too many open files
Here's my Camel route:
from("direct:publishCAT21FeatureCollection").process(
cat21FeatureCollectionToGMLTransformer).to(
"activemq-sender:topic:com.classactionpl.javaFlightTopic");
My spring config for AMQ is (activemq-receiver is defined but actually not
used by my app at the moment):
<bean id="activeMQConnectionFactory"
class="org.apache.activemq.ActiveMQConnectionFactory">
<property name="brokerURL"
value="${com.classactionpl.asterixservices.brokerurl}"
/>
</bean>
<bean id="activemq-receiver"
class="org.apache.activemq.camel.component.ActiveMQComponent">
<property name="connectionFactory"
ref="activeMQConnectionFactory"
/>
</bean>
<bean id="activemq-sender"
class="org.apache.activemq.camel.component.ActiveMQComponent">
<property name="connectionFactory">
<bean
class="org.apache.activemq.pool.PooledConnectionFactory">
<constructor-arg
ref="activeMQConnectionFactory" />
</bean>
</property>
</bean>
and the broker url:
com.classactionpl.asterixservices.brokerurl=failover:tcp://localhost:61616
Here's what I see in my Camel application's log file:
2009-07-09 09:47:09,607 [ActiveMQ Task ] INFO FailoverTransport
- Successfully reconnected to tcp://localhost:61616
2009-07-09 09:47:25,123 [127.0.0.1:61616] WARN FailoverTransport
- Transport failed to tcp://localhost:61616 , attempting to automatically
reconnect due to: java.net.SocketException: Connection reset
2009-07-09 09:47:25,132 [ActiveMQ Task ] INFO FailoverTransport
- Successfully reconnected to tcp://localhost:61616
2009-07-09 09:48:35,279 [127.0.0.1:61616] WARN FailoverTransport
- Transport failed to tcp://localhost:61616 , attempting to automatically
reconnect due to: java.io.EOFException
2009-07-09 09:48:35,291 [ActiveMQ Task ] INFO FailoverTransport
- Successfully reconnected to tcp://localhost:61616
2009-07-09 09:48:35,337 [ActiveMQ Task ] INFO FailoverTransport
- Successfully connected to tcp://localhost:61616
2009-07-09 09:49:35,478 [ActiveMQ Task ] INFO FailoverTransport
- Successfully connected to tcp://localhost:61616
2009-07-09 09:51:28,171 [ActiveMQ Task ] INFO FailoverTransport
- Successfully connected to tcp://localhost:61616
2009-07-09 09:52:55,163 [ActiveMQ Task ] INFO FailoverTransport
- Successfully connected to tcp://localhost:61616
I wouldn't expect so many failovers although it might just be the
PooledConnectionFactory filling up its pool of connections.
Any ideas?
Kind regards,
Christopher
--
View this message in context:
http://www.nabble.com/JMS-publish-consuming-sockets--tp24401653p24401653.html
Sent from the Camel - Users (activemq) mailing list archive at Nabble.com.
