[rabbitmq-discuss] Failing on restart and recovering queues

[Matthias Radestock]

On 28/11/11 19:20, Pierre Payet wrote:
> After a crash of the Rabbitmq server,

Do you know what happened? Did it run out of disk space by any chance?

> I tried to restart it but it get stuck on the step of recovering the
> different exchanges, queues and bindings. I have checked in
> /msg_store_persistent directory of the node and it seems the data are
> still here.

Right. It looks like some of the data got corrupted though. We have a 
bug open to make rabbit skip over such data on restart - we already do 
this elsewhere on recovery, but not in this particular case.

> Can anyone help me to recover the messages in the queue ?

Try deleting 
/var/lib/rabbitmq/mnesia/rabbit at curtiz/queues/67BZ7QWHYIY1S405V2E0881G9/37.idx. 
Hopefully that will allow rabbit to start with at least some of the data 
still present.


Regards,

Matthias.