[rabbitmq-discuss] Help pinpointing an error

[Tim Watson]

On 31 Aug 2012, at 12:32, Jaime Herazo B. wrote:

> I just checked on the machine with "rabbitmqctl list_queues name
> durable", and all show "false". Seems like someone here made a mistake
> and forgot to make the queues durable, which was something we all had
> assumed. Sorry for the alarm.
> 

Well at least the message loss is not unexpected then

> Is it possible to change the parameters of the queues and make them
> durable without having to make a new one with a different name as
> suggested in the Python tutorial?

Not that I'm aware of.

> 
> Also the broker still crashed, so i'm posting the info asked anyway:
> 
> * rabbitmq-server 2.8.4-1 on Debian Stable (6.0.3)

That sounds new enough to me - Matthias?

> * The service crashed, so i did a "service rabbitmq-server restart".
> Naturally, since now we know that the queues are not durable as we
> expected, all was lost.
> * No HA, no clustering, 2 machines sending information to each other,
> and a couple dozen clients consuming the messages put on the queues by
> the one that crashed.
> 
> And for the logs, here's the pastebin:
> http://pastebin.com/FTHGQVPQ
> The sasl log has far more information, but it was almost 1Mb of
> information on that crash alone, so initially i'm just pasting this and
> if needed i'll just add the other one.
> 

Is that it? It seems like you've pasted only a subset of the log file, and I'd like to know a bit more about what was happening on both nodes when this occurs. Can we see the sasl logs for both nodes too please, and the output of `rabbitmqctl report` for both nodes?