[rabbitmq-discuss] Rabbit MQ Crash

[Matthias Radestock]

On 07/04/14 21:22, Michael Sander wrote:
> Those timestamps seem accurate.  They are in UTC, so it was really
> around 6AM ET (where I am).  Those were in the sasl logs, I couldn't
> find anything interesting in the normal rabbitmq log at the same time.

Did you post the *entire* contents of the SASL log?

> Where else should I be looking for relevant logging info?

The sasl and normal log is all we need.

> Also, is this crash really due to the disk monitor? It seems like if
> there is a problem with the disk monitor, it shouldn't take down all of
> rabbitmq.

Quite, but the odd thing here is that this appears to be a transient 
error. Rabbit will disable disk monitoring when it fails on startup, 
whereas the situation here is that disk monitoring works initially and 
then fails at some point. When that happens rabbit will attempt to 
restart the disk monitor, and if that fails a few times *then* it will die.

We have some suspicion that the underlying cause might be a system 
update, in particular one that installs/updates Erlang. So it is worth 
checking that what cron jobs might be running on your system at 6am.

Matthias.