activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Claus Ibsen <claus.ib...@gmail.com>
Subject Re: [VOTE] Apache ActiveMQ 5.11.0 (rc3)
Date Wed, 04 Feb 2015 09:24:03 GMT
Hi

Just upgraded Camel master branch to 5.11.0 and you can reproduce the
issue there also

cd examples/camel-example-management
mvn clean install camel:run

Connect to the JVM using jconsole.

Notice the AcitveMQ tree under dynamic producers keeps adding new
mbeans, until your JVM runs out of memory.

On Wed, Feb 4, 2015 at 10:11 AM, Claus Ibsen <claus.ibsen@gmail.com> wrote:
> Hi
>
> I have done as Arthur suggested to dump the stacktrace, and also with
> Gary's turn on|off the anonoymous producers.
>
> I posted output from console as 2 gists
>
> https://gist.github.com/davsclaus/b5fbb1287d383107e599
>
> https://gist.github.com/davsclaus/0472ed0ab85ec5af9e79
>
> On Wed, Feb 4, 2015 at 12:05 AM, Gary Tully <gary.tully@gmail.com> wrote:
>> claus, that looks like producers are not being cached - I thought the
>> culprit may be https://issues.apache.org/jira/browse/AMQ-4968 but that
>> would have made 5.10. In any event, can you set
>> useAnonymousProducers=false on your activemq pooled connection
>> factory.
>>
>> On 3 February 2015 at 16:55, Claus Ibsen <claus.ibsen@gmail.com> wrote:
>>> Hi
>>>
>>> Just wanted to say I have stumbled on a weird issue with the 5.11.0
>>> release which I had trouble reproducing consistently.
>>>
>>> The issue is subtle causing AMQ to continuously re-create dynamic
>>> producer mbeans in the JMX tree. This is using a Camel spring app with
>>> the AMQ connection pool.
>>>
>>> Though just now I was able to see that 5.10.0 was stable and the
>>> mbeans didnt re-create continusly, but with 5.11.0 that happens (with
>>> no other change that switching from 5.10.0 to 5.11.0)
>>>
>>> Though I think that issue could go unnoticed as the jmx tree looks
>>> okay, but you may discover a flicker in jconsole when it redraws, and
>>> then there is a little CPU time to create/destroy the mbeans.
>>>
>>> I noticed this with hawtio as it also "flicker" when the jmx tree is
>>> redrawn. And therefore was using hawtio to track down the issue.
>>>
>>>
>>> I enabled a WARN logging when a mbean was registerede/unregistered and
>>> you can see
>>>
>>> WARN  | ActiveMQ Transport: tcp:///127.0.0.1:56632@31316 | JMX
>>> notification: javax.management.MBeanServerNotification[source=JMImplementation:type=MBeanServerDelegate][type=JMX.mbean.registered][message=][mbeanName=org.apache.activemq:type=Broker,brokerName=broker1,endpoint=dynamicProducer,clientId=ID_davsclaus.air-56625-1422981975940-4_5,producerId=ID_davsclaus.air-56625-1422981975940-5_6_1_8]
>>> WARN  | ActiveMQ Transport: tcp:///127.0.0.1:56633@31316 | JMX
>>> notification: javax.management.MBeanServerNotification[source=JMImplementation:type=MBeanServerDelegate][type=JMX.mbean.registered][message=][mbeanName=org.apache.activemq:type=Broker,brokerName=broker1,endpoint=dynamicProducer,clientId=ID_davsclaus.air-56625-1422981975940-4_6,producerId=ID_davsclaus.air-56625-1422981975940-5_7_1_8]
>>> WARN  | ActiveMQ Transport: tcp:///127.0.0.1:56634@31316 | JMX
>>> notification: javax.management.MBeanServerNotification[source=JMImplementation:type=MBeanServerDelegate][type=JMX.mbean.registered][message=][mbeanName=org.apache.activemq:type=Broker,brokerName=broker1,endpoint=dynamicProducer,clientId=ID_davsclaus.air-56625-1422981975940-4_7,producerId=ID_davsclaus.air-56625-1422981975940-5_8_1_8]
>>> WARN  | ActiveMQ Transport: tcp:///127.0.0.1:56626@31316 | JMX
>>> notification: javax.management.MBeanServerNotification[source=JMImplementation:type=MBeanServerDelegate][type=JMX.mbean.registered][message=][mbeanName=org.apache.activemq:type=Broker,brokerName=broker1,endpoint=dynamicProducer,clientId=ID_davsclaus.air-56625-1422981975940-4_8,producerId=ID_davsclaus.air-56625-1422981975940-5_1_1_8]
>>> WARN  | ActiveMQ Transport: tcp:///127.0.0.1:56628@31316 | JMX
>>> notification: javax.management.MBeanServerNotification[source=JMImplementation:type=MBeanServerDelegate][type=JMX.mbean.registered][message=][mbeanName=org.apache.activemq:type=Broker,brokerName=broker1,endpoint=dynamicProducer,clientId=ID_davsclaus.air-56625-1422981975940-4_1,producerId=ID_davsclaus.air-56625-1422981975940-5_2_11_8]
>>> WARN  | ActiveMQ Transport: tcp:///127.0.0.1:56629@31316 | JMX
>>> notification: javax.management.MBeanServerNotification[source=JMImplementation:type=MBeanServerDelegate][type=JMX.mbean.registered][message=][mbeanName=org.apache.activemq:type=Broker,brokerName=broker1,endpoint=dynamicProducer,clientId=ID_davsclaus.air-56625-1422981975940-4_2,producerId=ID_davsclaus.air-56625-1422981975940-5_3_1_9]
>>>
>>>
>>>
>>> Just wanted to share this and say there is maybe a glitch in this release.
>>>
>>> As I discovered this using hawtio, I created a branch with my current
>>> work code that reproduces the issue.
>>>
>>> https://github.com/hawtio/hawtio/tree/amq511bug
>>>
>>> The commit has instructions how to reproduce.
>>> https://github.com/hawtio/hawtio/commit/9162529299059d25e1e6b7c7d7e0be45ee4556b4
>>>
>>> As I was a hunt in the hay-stack I had to disable various stuff in the
>>> hawtio spring xml file where we setup stuff we use for testing and
>>> developing, hence the removal of the NOB, and some Camel routes and
>>> whatnot.
>>>
>>> On Fri, Jan 30, 2015 at 3:00 PM, Gary Tully <gary.tully@gmail.com> wrote:
>>>> Hi folks,
>>>>
>>>> I've just cut a third release candidate for the long-awaited 5.11.0 release.
>>>> This release has more than 120 bug fixes and improvements.
>>>>
>>>> -note-
>>>> this candidate includes the fix for the 'reliably' broken test case from
rc2
>>>> however there can be no expectation of a 'reliable' full test run from
>>>> mvn install.
>>>> I have raised https://issues.apache.org/jira/browse/AMQ-5552 to address that.
>>>> I think a resolution to that issue would address all of the
>>>> disapproving commentary on the rc2 vote. If you voted -1 last time
>>>> round, please chime in on AMQ-5552.
>>>> -end note-
>>>>
>>>> Could you please review the artifacts and vote?
>>>>
>>>> The list of resolved issues is here:
>>>> https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12311210&version=12324951
>>>>
>>>> You can get binary distributions here:
>>>> https://repository.apache.org/content/repositories/orgapacheactivemq-1016/org/apache/activemq/apache-activemq/5.11.0/
>>>>
>>>> Source archives are here:
>>>> https://repository.apache.org/content/repositories/orgapacheactivemq-1016/org/apache/activemq/activemq-parent/5.11.0/
>>>>
>>>> Maven2 repository is at:
>>>> https://repository.apache.org/content/repositories/orgapacheactivemq-1016/
>>>>
>>>> Source tag:
>>>> https://git-wip-us.apache.org/repos/asf?p=activemq.git;a=tag;h=66bd5da9405765ddbc950b4f1c5a6ef6195207e5
>>>>
>>>> The vote will remain open for 72 hours.
>>>>
>>>> [ ] +1  approve the release as Apache ActiveMQ 5.11.0
>>>> [ ] +0  no opinion
>>>> [ ] -1  disapprove (and reason why)
>>>>
>>>>
>>>> Here's my +1
>>>>
>>>> Regards,
>>>> Gary.
>>>
>>>
>>>
>>> --
>>> Claus Ibsen
>>> -----------------
>>> Red Hat, Inc.
>>> Email: cibsen@redhat.com
>>> Twitter: davsclaus
>>> Blog: http://davsclaus.com
>>> Author of Camel in Action: http://www.manning.com/ibsen
>>> hawtio: http://hawt.io/
>>> fabric8: http://fabric8.io/
>
>
>
> --
> Claus Ibsen
> -----------------
> Red Hat, Inc.
> Email: cibsen@redhat.com
> Twitter: davsclaus
> Blog: http://davsclaus.com
> Author of Camel in Action: http://www.manning.com/ibsen
> hawtio: http://hawt.io/
> fabric8: http://fabric8.io/



-- 
Claus Ibsen
-----------------
Red Hat, Inc.
Email: cibsen@redhat.com
Twitter: davsclaus
Blog: http://davsclaus.com
Author of Camel in Action: http://www.manning.com/ibsen
hawtio: http://hawt.io/
fabric8: http://fabric8.io/

Mime
View raw message