You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Torbjorn Kristoffersen <to...@gmail.com> on 2010/09/25 23:23:33 UTC
Dispatch problems in a simple store and forward setup with ActiveMQ 5.4.1
I have two ActiveMQ brokers in a simple store and forward setup.
* Broker B1 is behind a firewall, so it's set up with a duplex
connection to the other broker, B2 (with a public IP).
* Consumer CCC is connected to B2 and has a durable subscription to TestTopic.
Here is the flow of events:
1) Someone on B1's network sends a message for TestTopic
2) The consumer CCC on B2, gets the message since the store and
forwarding does its job.
Until this point, everything works fine. The problems begin when the
connection between B1 and B2 goes down:
3) The connection between B1 and B2 goes down
4) Someone on B1's network sends another message to TestTopic
5) B1 hangs around, waiting for the connection to come back up. (zzzZzz)
6) After a few minutes, the internet is working again. Duplex
connection is back up!
7) The message goes to B2, but the consumer CCC gets absolutely nothing.
Only when CCC re-connects to B2, he gets the message. Why does
this happen? This consumer can't predict when it should reconnect
because
it had no idea that the link to B1 went down)
It is also worth mentioning that I am using ActiveMQ 5.4.1 (also tried
with 5.3.2). The consumer CCC is connected via STOMP.
Thank you
Re: Dispatch problems in a simple store and forward setup with
ActiveMQ 5.4.1
Posted by Dejan Bosanac <de...@nighttale.net>.
Hi,
the more the merrier.
client, configuration and explanation on how to reproduce it would be the best
Cheers
--
Dejan Bosanac - http://twitter.com/dejanb
Open Source Integration - http://fusesource.com/
ActiveMQ in Action - http://www.manning.com/snyder/
Blog - http://www.nighttale.net
On Mon, Sep 27, 2010 at 2:21 PM, Torbjorn Kristoffersen
<to...@gmail.com> wrote:
> Hi Dejan
>
> By test case do you think it's best to include:
>
> 1) Both activemq config files
> 2) Client STOMP program (a little python client)
>
> Or should it just be the whole log and detailed explanations.
>
> On Mon, Sep 27, 2010 at 10:00 AM, Dejan Bosanac <de...@nighttale.net> wrote:
>> Hi Torbjorn,
>>
>> can you raise a Jira issue for this, ideally with the test case?
>>
>> Cheers
>> --
>> Dejan Bosanac - http://twitter.com/dejanb
>>
>> Open Source Integration - http://fusesource.com/
>> ActiveMQ in Action - http://www.manning.com/snyder/
>> Blog - http://www.nighttale.net
>>
>>
>>
>> On Sat, Sep 25, 2010 at 11:23 PM, Torbjorn Kristoffersen
>> <to...@gmail.com> wrote:
>>> I have two ActiveMQ brokers in a simple store and forward setup.
>>>
>>> * Broker B1 is behind a firewall, so it's set up with a duplex
>>> connection to the other broker, B2 (with a public IP).
>>>
>>> * Consumer CCC is connected to B2 and has a durable subscription to TestTopic.
>>>
>>> Here is the flow of events:
>>>
>>> 1) Someone on B1's network sends a message for TestTopic
>>> 2) The consumer CCC on B2, gets the message since the store and
>>> forwarding does its job.
>>>
>>> Until this point, everything works fine. The problems begin when the
>>> connection between B1 and B2 goes down:
>>>
>>> 3) The connection between B1 and B2 goes down
>>> 4) Someone on B1's network sends another message to TestTopic
>>> 5) B1 hangs around, waiting for the connection to come back up. (zzzZzz)
>>> 6) After a few minutes, the internet is working again. Duplex
>>> connection is back up!
>>> 7) The message goes to B2, but the consumer CCC gets absolutely nothing.
>>> Only when CCC re-connects to B2, he gets the message. Why does
>>> this happen? This consumer can't predict when it should reconnect
>>> because
>>> it had no idea that the link to B1 went down)
>>>
>>> It is also worth mentioning that I am using ActiveMQ 5.4.1 (also tried
>>> with 5.3.2). The consumer CCC is connected via STOMP.
>>>
>>> Thank you
>>>
>>
>
Re: Dispatch problems in a simple store and forward setup with
ActiveMQ 5.4.1
Posted by Torbjorn Kristoffersen <to...@gmail.com>.
Hi Dejan
By test case do you think it's best to include:
1) Both activemq config files
2) Client STOMP program (a little python client)
Or should it just be the whole log and detailed explanations.
On Mon, Sep 27, 2010 at 10:00 AM, Dejan Bosanac <de...@nighttale.net> wrote:
> Hi Torbjorn,
>
> can you raise a Jira issue for this, ideally with the test case?
>
> Cheers
> --
> Dejan Bosanac - http://twitter.com/dejanb
>
> Open Source Integration - http://fusesource.com/
> ActiveMQ in Action - http://www.manning.com/snyder/
> Blog - http://www.nighttale.net
>
>
>
> On Sat, Sep 25, 2010 at 11:23 PM, Torbjorn Kristoffersen
> <to...@gmail.com> wrote:
>> I have two ActiveMQ brokers in a simple store and forward setup.
>>
>> * Broker B1 is behind a firewall, so it's set up with a duplex
>> connection to the other broker, B2 (with a public IP).
>>
>> * Consumer CCC is connected to B2 and has a durable subscription to TestTopic.
>>
>> Here is the flow of events:
>>
>> 1) Someone on B1's network sends a message for TestTopic
>> 2) The consumer CCC on B2, gets the message since the store and
>> forwarding does its job.
>>
>> Until this point, everything works fine. The problems begin when the
>> connection between B1 and B2 goes down:
>>
>> 3) The connection between B1 and B2 goes down
>> 4) Someone on B1's network sends another message to TestTopic
>> 5) B1 hangs around, waiting for the connection to come back up. (zzzZzz)
>> 6) After a few minutes, the internet is working again. Duplex
>> connection is back up!
>> 7) The message goes to B2, but the consumer CCC gets absolutely nothing.
>> Only when CCC re-connects to B2, he gets the message. Why does
>> this happen? This consumer can't predict when it should reconnect
>> because
>> it had no idea that the link to B1 went down)
>>
>> It is also worth mentioning that I am using ActiveMQ 5.4.1 (also tried
>> with 5.3.2). The consumer CCC is connected via STOMP.
>>
>> Thank you
>>
>
Re: Dispatch problems in a simple store and forward setup with
ActiveMQ 5.4.1
Posted by Dejan Bosanac <de...@nighttale.net>.
Hi Torbjorn,
can you raise a Jira issue for this, ideally with the test case?
Cheers
--
Dejan Bosanac - http://twitter.com/dejanb
Open Source Integration - http://fusesource.com/
ActiveMQ in Action - http://www.manning.com/snyder/
Blog - http://www.nighttale.net
On Sat, Sep 25, 2010 at 11:23 PM, Torbjorn Kristoffersen
<to...@gmail.com> wrote:
> I have two ActiveMQ brokers in a simple store and forward setup.
>
> * Broker B1 is behind a firewall, so it's set up with a duplex
> connection to the other broker, B2 (with a public IP).
>
> * Consumer CCC is connected to B2 and has a durable subscription to TestTopic.
>
> Here is the flow of events:
>
> 1) Someone on B1's network sends a message for TestTopic
> 2) The consumer CCC on B2, gets the message since the store and
> forwarding does its job.
>
> Until this point, everything works fine. The problems begin when the
> connection between B1 and B2 goes down:
>
> 3) The connection between B1 and B2 goes down
> 4) Someone on B1's network sends another message to TestTopic
> 5) B1 hangs around, waiting for the connection to come back up. (zzzZzz)
> 6) After a few minutes, the internet is working again. Duplex
> connection is back up!
> 7) The message goes to B2, but the consumer CCC gets absolutely nothing.
> Only when CCC re-connects to B2, he gets the message. Why does
> this happen? This consumer can't predict when it should reconnect
> because
> it had no idea that the link to B1 went down)
>
> It is also worth mentioning that I am using ActiveMQ 5.4.1 (also tried
> with 5.3.2). The consumer CCC is connected via STOMP.
>
> Thank you
>