[rabbitmq-discuss] server crashes with very fast consumers

alex chen chen650 at yahoo.com
Thu Mar 17 05:39:20 GMT 2011


I tried building my previous patched amqp_consumer with latest rabbitmq-c, 
however, the build breaks because the current amqp.h does not support 
AMQP_SIMPLE_RPC anymore.  Ideally amqp.h should support the following function, 
as commented in my patch:

/* amqp.h should support this */
int amqp_basic_qos(amqp_connection_state_t state,
                           amqp_channel_t channel,
                           uint32_t prefetch_size,
                           uint16_t prefetch_count,
                           amqp_boolean_t global)

It would be nice if rabbitmq-c maintainer could add this api.  otherwise, i have 
to take a look at the src code and see if it can be implemented.



From: John DeTreville <john at detreville.org>
To: alex chen <chen650 at yahoo.com>
Cc: rabbitmq <rabbitmq-discuss at lists.rabbitmq.com>
Sent: Wed, March 16, 2011 7:04:08 PM
Subject: Re: [rabbitmq-discuss] server crashes with very fast consumers


I'm trying to reproduce your problems and I'm having a very hard time doing so.

Your example code is based on your own modifications to the sources in the 
rabbitmq-c/examples directory in some revision from last summer, which itself 
depends on some revision of rabbitmq-codegen from last summer. I'm having a hard 
time building it.

I'm guessing you've run all of this much more recently. Could you please mail us 
a bind.sh, producer.sh, and consumer.sh that work with the current 
rabbitmq-c/examples directory? Could you also add your modifications to the 
source code in the current rabbitmq-c/examples directory? This would help a 
great deal.

Even so, I've been able to reproduce some crashes, but I very much doubt they're 
the same crashes that you're seeing. Again, please tell me exactly what to do to 
reproduce your problem, starting with the current version of rabbitmq-c. Thank 


On Mar 14, 2011, at 4:29 PM, Matthias Radestock wrote:

> Alex,
> alex chen wrote:
>>> Any chance you could script the above to the point where all one 
>> needs to do is:
>>> 1) start clean rabbit
>>> 2) run test script
>>> 3) wait for broker to die
>> The test scripts are exactly the same as before:
>> http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/2010-August/008570.html
> It would really help if you consolidated all that into a simple set up 
> as I mention above. As it stands we have to piece together the test code 
> and scripts from various emails. That a) takes time, and b) increases 
> the chances that we will be doing something slightly different and thus 
> not be able to reproduce the problem.
> Regards,
> Matthias.
> PS: please keep rabbitmq-discuss on the cc.
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20110316/de4cf3d5/attachment.htm>

More information about the rabbitmq-discuss mailing list