[rabbitmq-discuss] Failing on restart and recovering queues

[Matthias Radestock]

On 28/11/11 21:13, Pierre Payet wrote:
> It's not the disk space but I'm not sure of how to answer this question.
> Looking at the rabbitmq at curtiz.log.1, the last event reported is a tcp
> connection attempt but no trace of the fulfilling of it.
> Looking at rabbitmq-sasl.log.1 I just have this message :
> =SUPERVISOR REPORT==== 23-Nov-2011::17:36:29 ===
>       Supervisor: {<0.1220.149>,
>                                             rabbit_channel_sup_sup}
>       Context:    shutdown_error
>       Reason:     shutdown
>       Offender:   [{pid,<0.1222.149>},
>                    {name,channel_sup},
>                    {mfa,{rabbit_channel_sup,start_link,[]}},
>                    {restart_type,temporary},
>                    {shutdown,infinity},
>                    {child_type,supervisor}]

Are you sure that is the last event logged by the crashed rabbit? In 
recent rabbit releases we no longer move log contents to the .1 file on 
startup. Instead we simply append to the existing file. So perhaps there 
are errors in the rabbit at curtiz-sasl.log relating to the crash? Also, 
did you get an erl_crash.dump?

> I have remarked that i have a lot's of connection running at the same
> time. Could it be a too high amount of connection to the server that
> crashed it ?

The server guards against that; it should refuse connection attempts 
when file descriptors get tight.

Regards,

Matthias.