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

[Thiago Burgo Belo]

The channels (~1200) was created by a problem in my tests with a
IConnection Pool, where I opened a channel (CreateModel) and I forgot
dispose.

You know when a new version with this bug fix will be avaiable?

Thanks for help,

2012/7/17 Matthias Radestock <matthias at rabbitmq.com>

> Thiago,
>
> On 17/07/12 16:09, Thiago Burgo Belo wrote:
>
>> https://skydrive.live.com/**redir?resid=E7129E57616CE33A!**
>> 4492&authkey=!AB9nHGtNRRUeC94<https://skydrive.live.com/redir?resid=E7129E57616CE33A%214492&authkey=%21AB9nHGtNRRUeC94>
>>
>
> Thank you.
>
> We think we know what the problem is: a race condition when repeatedly
> declaring and deleting the same HA queue. This got fixed on 'default' a
> couple of weeks ago and will appear in the next release.
>
> Btw, one oddity I spotted in your logs is that at one point a queue
> appeared to have ~1200 channels that each had published a message to it and
> where all still alive. That may be perfectly ok of course, but it is quite
> unusual.
>
> Regards,
>
> Matthias.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20120717/0f63322c/attachment.htm>