[rabbitmq-discuss] My cluter node (ram node) down in my performance test

Thiago Burgo Belo thiagoburgo at gmail.com
Fri Jul 6 15:15:16 BST 2012


I delete my log file (~200MB) =(, I will try reproduce problem (the problem
happens frequently), and after a send the log for you. Wait a 30 min to my
performance test run, in my test I simulate 100k messages sent by 60
concurrent users (using apache benchmark -ab.exe).

Thiago Burgo

2012/7/6 Matthias Radestock <matthias at rabbitmq.com>

> Thiago,
> On 06/07/12 14:21, Thiago Burgo Belo wrote:
>> I'm doing performance tests with Rabbit 2.8.4 in cluter mode (2 nodes),
>> when I send 100000 (2 publishers 50K messages) messages with 1 client
>> (consumer) with .NET client, my "slave" node running at ram down suddenly.
>> Other problem (I dont know if correlated) is I lost messages, about 8%
>> of loss.
>> In attachment the server log.
> That's some interesting errors you've got there. Could you please send us
> - the *complete* log file (if it's huge just put it somewhere where we can
> download it)
> - the output of 'rabbitmqctl report' on the ram node when your tests are
> running (but before the crash)
> Regards,
> Matthias.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20120706/3c7d0c15/attachment.htm>

More information about the rabbitmq-discuss mailing list