[rabbitmq-discuss] SSL connection errors once or twice every week

[Simon MacMullen]

On 14/03/12 00:31, Brian Jing wrote:
> We are currently using non-ssl ports, and only one-way federation in our
> system. It seems that the whole thing is more stable now. However, the
> hanging still happened last weekend. Is there anything from our end that
> could help you pinpoint what the problem is? Logs or coredumps?

For the record, symmetric federation should work fine, it's just 
inefficient with bandwidth (this will change).

Does anything interesting appear in the logs when you try to restart and 
it hangs? Or at any other time?

Earlier in the thread we saw ehostunreach errors on running connections 
- i.e. a host we had a working TCP connection to stopped being routable. 
I'm not even sure how that would happen - is there anything unusual in 
your network that would explain it?

Also, do you have heartbeats turned on?

Cheers, Simon

-- 
Simon MacMullen
RabbitMQ, VMware