You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Dhruba Bandopadhyay <bd...@gmail.com> on 2008/04/17 15:24:10 UTC

ActiveMQ-5.1.0-RC3 - Messages going missing

Hi,

Using spring 2.5.3 mdps, tomcat 6.0.16 and activemq 5.1.0-rc3 with 1
producer, 3 brokers and 1 consumer running 5 threads with almost
default broker configuration.

Producer and consumer connecting via the following url.

failover://(tcp://10.10.32.140:61616,tcp://10.10.32.142:61616,tcp://10.10.32.179:61616)?initialReconnectDelay=100&randomize=false

All brokers using following config.

<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:amq="http://activemq.org/config/1.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans-2.0.xsd
  http://activemq.org/config/1.0
http://activemq.apache.org/schema/activemq-core.xsd
  http://activemq.apache.org/camel/schema/spring
http://activemq.apache.org/camel/schema/spring/camel-spring.xsd">

    <bean class="org.springframework.beans.factory.config.PropertyPlaceholderConfigurer"
/>

    <broker xmlns="http://activemq.org/config/1.0"
brokerName="broker3" dataDirectory="${activemq.base}/data">
        <transportConnectors>
            <transportConnector name="openwire" uri="tcp://localhost:61616" />
        </transportConnectors>
    </broker>

    <jetty xmlns="http://mortbay.com/schemas/jetty/1.0">
        <connectors>
            <nioConnector port="8161" />
        </connectors>
        <handlers>
            <webAppContext contextPath="/admin"
resourceBase="${activemq.base}/webapps/admin" logUrlOnStart="true" />
            <webAppContext contextPath="/demo"
resourceBase="${activemq.base}/webapps/demo" logUrlOnStart="true" />
            <webAppContext contextPath="/fileserver"
resourceBase="${activemq.base}/webapps/fileserver"
logUrlOnStart="true" />
        </handlers>
    </jetty>

</beans>

Due to the client connection url the single consumer only uses the
first broker.  Note all multicast options are disabled to prevent
messages going from one broker to another.

When I send 20,000 messages frequently roughly 1000 or so go missing.

Is this correct behaviour given my setup?  What are the usual causes
of messages going missing?  I don't see any exceptions or failures
anywhere.  The consumer stats show that the consumer has only
processed roughly 19000 messages so it is definitely not the case that
it is has processed them but failed to do so correctly.

Any help, advice or link to appropriate documentation would be much
appreciated as for me this is a serious problem.

Many thanks.

Re: ActiveMQ-5.1.0-RC3 - Messages going missing

Posted by alanmc <al...@alum.mit.edu>.
Sorry for the long delay responding.  I resolved this issue, it was a bug in
our client code.  A temporary queue was being cleaned up too early.


rajdavies wrote:
> 
> Are you using networks as well ?
> 
> On 22 Apr 2008, at 04:01, alanmc wrote:
> 
>>
>> We're seeing the same behavior with 5.1.0.  We're basically running  
>> the
>> default broker configuration, with advisory support turned off.  We  
>> were
>> using 5.0.0, but saw memory leakage and wanted to try out 5.1.0 and  
>> after
>> switching over with the exact same configuration we started seeing
>> intermittent message dropping, probably about 1 in 3 messages going  
>> missing.
>> The broker shows all messages as enqueued and dequeued.
>>
>> We're using ActiveMQ as middleware between a website and back-end  
>> services.
>> We run python clients and model Request-Reply by creating a temp  
>> queue and
>> sending the request with the reply-to attribute populated.  Works  
>> great on
>> 5.0.0 (except for troubling memory usage), but switch to 5.1.0 and 1  
>> in 3
>> messages just disappear.
>>
>>
>> Dhruba Bandopadhyay-2 wrote:
>>>
>>> Further to this I'd like to add that the broker shows that all 20000
>>> messages have been enqueued and dequeued.  So the broker queue has no
>>> messages remaining on it.
>>>
>>> On Thu, Apr 17, 2008 at 2:24 PM, Dhruba Bandopadhyay <bdhruba@gmail.com 
>>> >
>>> wrote:
>>>> Hi,
>>>>
>>>> Using spring 2.5.3 mdps, tomcat 6.0.16 and activemq 5.1.0-rc3 with 1
>>>> producer, 3 brokers and 1 consumer running 5 threads with almost
>>>> default broker configuration.
>>>
>>>> Is this correct behaviour given my setup?  What are the usual causes
>>>> of messages going missing?  I don't see any exceptions or failures
>>>> anywhere.  The consumer stats show that the consumer has only
>>>> processed roughly 19000 messages so it is definitely not the case  
>>>> that
>>>> it is has processed them but failed to do so correctly.
>>>>
>>>> Any help, advice or link to appropriate documentation would be much
>>>> appreciated as for me this is a serious problem.
>>>>
>>>> Many thanks.
>>>>
>>>
>>>
>>
>> -- 
>> View this message in context:
>> http://www.nabble.com/ActiveMQ-5.1.0-RC3---Messages-going-missing-tp16746451s2354p16820504.html
>> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>>
> 
> 
> 

-- 
View this message in context: http://www.nabble.com/ActiveMQ-5.1.0-RC3---Messages-going-missing-tp16746451p18139017.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.


Re: ActiveMQ-5.1.0-RC3 - Messages going missing

Posted by Rob Davies <ra...@gmail.com>.
Are you using networks as well ?

On 22 Apr 2008, at 04:01, alanmc wrote:

>
> We're seeing the same behavior with 5.1.0.  We're basically running  
> the
> default broker configuration, with advisory support turned off.  We  
> were
> using 5.0.0, but saw memory leakage and wanted to try out 5.1.0 and  
> after
> switching over with the exact same configuration we started seeing
> intermittent message dropping, probably about 1 in 3 messages going  
> missing.
> The broker shows all messages as enqueued and dequeued.
>
> We're using ActiveMQ as middleware between a website and back-end  
> services.
> We run python clients and model Request-Reply by creating a temp  
> queue and
> sending the request with the reply-to attribute populated.  Works  
> great on
> 5.0.0 (except for troubling memory usage), but switch to 5.1.0 and 1  
> in 3
> messages just disappear.
>
>
> Dhruba Bandopadhyay-2 wrote:
>>
>> Further to this I'd like to add that the broker shows that all 20000
>> messages have been enqueued and dequeued.  So the broker queue has no
>> messages remaining on it.
>>
>> On Thu, Apr 17, 2008 at 2:24 PM, Dhruba Bandopadhyay <bdhruba@gmail.com 
>> >
>> wrote:
>>> Hi,
>>>
>>> Using spring 2.5.3 mdps, tomcat 6.0.16 and activemq 5.1.0-rc3 with 1
>>> producer, 3 brokers and 1 consumer running 5 threads with almost
>>> default broker configuration.
>>
>>> Is this correct behaviour given my setup?  What are the usual causes
>>> of messages going missing?  I don't see any exceptions or failures
>>> anywhere.  The consumer stats show that the consumer has only
>>> processed roughly 19000 messages so it is definitely not the case  
>>> that
>>> it is has processed them but failed to do so correctly.
>>>
>>> Any help, advice or link to appropriate documentation would be much
>>> appreciated as for me this is a serious problem.
>>>
>>> Many thanks.
>>>
>>
>>
>
> -- 
> View this message in context: http://www.nabble.com/ActiveMQ-5.1.0-RC3---Messages-going-missing-tp16746451s2354p16820504.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>


Re: ActiveMQ-5.1.0-RC3 - Messages going missing

Posted by alanmc <al...@alum.mit.edu>.
We're seeing the same behavior with 5.1.0.  We're basically running the
default broker configuration, with advisory support turned off.  We were
using 5.0.0, but saw memory leakage and wanted to try out 5.1.0 and after
switching over with the exact same configuration we started seeing
intermittent message dropping, probably about 1 in 3 messages going missing. 
The broker shows all messages as enqueued and dequeued.

We're using ActiveMQ as middleware between a website and back-end services. 
We run python clients and model Request-Reply by creating a temp queue and
sending the request with the reply-to attribute populated.  Works great on
5.0.0 (except for troubling memory usage), but switch to 5.1.0 and 1 in 3
messages just disappear.


Dhruba Bandopadhyay-2 wrote:
> 
> Further to this I'd like to add that the broker shows that all 20000
> messages have been enqueued and dequeued.  So the broker queue has no
> messages remaining on it.
> 
> On Thu, Apr 17, 2008 at 2:24 PM, Dhruba Bandopadhyay <bd...@gmail.com>
> wrote:
>> Hi,
>>
>>  Using spring 2.5.3 mdps, tomcat 6.0.16 and activemq 5.1.0-rc3 with 1
>>  producer, 3 brokers and 1 consumer running 5 threads with almost
>>  default broker configuration.
> 
>>  Is this correct behaviour given my setup?  What are the usual causes
>>  of messages going missing?  I don't see any exceptions or failures
>>  anywhere.  The consumer stats show that the consumer has only
>>  processed roughly 19000 messages so it is definitely not the case that
>>  it is has processed them but failed to do so correctly.
>>
>>  Any help, advice or link to appropriate documentation would be much
>>  appreciated as for me this is a serious problem.
>>
>>  Many thanks.
>>
> 
> 

-- 
View this message in context: http://www.nabble.com/ActiveMQ-5.1.0-RC3---Messages-going-missing-tp16746451s2354p16820504.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.


Re: ActiveMQ-5.1.0-RC3 - Messages going missing

Posted by Dhruba Bandopadhyay <bd...@gmail.com>.
Further to this I'd like to add that the broker shows that all 20000
messages have been enqueued and dequeued.  So the broker queue has no
messages remaining on it.

On Thu, Apr 17, 2008 at 2:24 PM, Dhruba Bandopadhyay <bd...@gmail.com> wrote:
> Hi,
>
>  Using spring 2.5.3 mdps, tomcat 6.0.16 and activemq 5.1.0-rc3 with 1
>  producer, 3 brokers and 1 consumer running 5 threads with almost
>  default broker configuration.
>
>  Producer and consumer connecting via the following url.
>
>  failover://(tcp://10.10.32.140:61616,tcp://10.10.32.142:61616,tcp://10.10.32.179:61616)?initialReconnectDelay=100&randomize=false
>
>  All brokers using following config.
>
>  <beans xmlns="http://www.springframework.org/schema/beans"
>  xmlns:amq="http://activemq.org/config/1.0"
>  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
>     xsi:schemaLocation="http://www.springframework.org/schema/beans
>  http://www.springframework.org/schema/beans/spring-beans-2.0.xsd
>   http://activemq.org/config/1.0
>  http://activemq.apache.org/schema/activemq-core.xsd
>   http://activemq.apache.org/camel/schema/spring
>  http://activemq.apache.org/camel/schema/spring/camel-spring.xsd">
>
>     <bean class="org.springframework.beans.factory.config.PropertyPlaceholderConfigurer"
>  />
>
>     <broker xmlns="http://activemq.org/config/1.0"
>  brokerName="broker3" dataDirectory="${activemq.base}/data">
>         <transportConnectors>
>             <transportConnector name="openwire" uri="tcp://localhost:61616" />
>         </transportConnectors>
>     </broker>
>
>     <jetty xmlns="http://mortbay.com/schemas/jetty/1.0">
>         <connectors>
>             <nioConnector port="8161" />
>         </connectors>
>         <handlers>
>             <webAppContext contextPath="/admin"
>  resourceBase="${activemq.base}/webapps/admin" logUrlOnStart="true" />
>             <webAppContext contextPath="/demo"
>  resourceBase="${activemq.base}/webapps/demo" logUrlOnStart="true" />
>             <webAppContext contextPath="/fileserver"
>  resourceBase="${activemq.base}/webapps/fileserver"
>  logUrlOnStart="true" />
>         </handlers>
>     </jetty>
>
>  </beans>
>
>  Due to the client connection url the single consumer only uses the
>  first broker.  Note all multicast options are disabled to prevent
>  messages going from one broker to another.
>
>  When I send 20,000 messages frequently roughly 1000 or so go missing.
>
>  Is this correct behaviour given my setup?  What are the usual causes
>  of messages going missing?  I don't see any exceptions or failures
>  anywhere.  The consumer stats show that the consumer has only
>  processed roughly 19000 messages so it is definitely not the case that
>  it is has processed them but failed to do so correctly.
>
>  Any help, advice or link to appropriate documentation would be much
>  appreciated as for me this is a serious problem.
>
>  Many thanks.
>