[rabbitmq-discuss] 3.02 numerous log error reports

Anton Lebedevich mabrek at gmail.com
Mon Feb 25 11:46:19 GMT 2013


On 02/25/2013 05:29 AM, Matthias Radestock wrote:
> Steve,
> 
> On 25/02/13 00:45, steve flitcroft wrote:
>> Forgot to also mention that I am running haproxy in front of the rmq
>> node.
> 
> haproxy is probably causing the errors...
> 
>> =ERROR REPORT==== 25-Feb-2013::00:22:00 === ** Generic server
>> <0.26372.31> terminating ** Last message in was
>> {inet_async,#Port<0.17264>,6622,{error,econnaborted}} ** When Server
>> state == {state,{rabbit_networking,start_client,[]},
>> #Port<0.17264>,6622} ** Reason for termination == **
>> {accept_failed,econnaborted}
> 
> That error indicates that an inbound connection got closed before rabbit
> got a chance to handle it.
> 
>> =ERROR REPORT==== 25-Feb-2013::00:22:05 === error on AMQP connection
>> <0.26419.31>: enotconn
> 
> That error indicates that an inbound connection got closed just after
> rabbit accepted it.
> 
> Both errors are harmless and rabbit should probably just ignore them, or
> log something less frightening.
> 
> 
> The reason you are seeing this error is probably due to haproxy trying
> to determine whether a node is alive by attempting to connect to
> the AMQP port and closing the connection immediately.
> 
> It's curious though that no user has reported this before.

Is it possible to check rabbitmq aliveness from haproxy without spamming
logs with these error reports?

Haproxy can do http checks, but it can't supply user/password needed for
api/aliveness-test.

Regards,
Anton Lebedevich.


More information about the rabbitmq-discuss mailing list