[rabbitmq-discuss] memory leak rabbitmq 3.1.5
Simon MacMullen
simon at rabbitmq.com
Thu Sep 26 15:42:17 BST 2013
Ah, thank you. Sounds like the cause. I'll file a bug; in the mean time
I suggest you only use 5672 for keepalived.
Cheers, Simon
On 26/09/2013 3:27PM, Santi Nuñez-Cacho wrote:
> ok, I got keepalived above the cluster. And is balancing 61613 and 5672
> could be keepalived opening and not closing something properly?
> I undertstand the connections are being close well, but ..
>
>
> On Thu, Sep 26, 2013 at 4:18 PM, Santi Nuñez-Cacho
> <brian.santiago at gmail.com <mailto:brian.santiago at gmail.com>> wrote:
>
> yes, I do
>
> =INFO REPORT==== 26-Sep-2013::16:12:46 ===
> accepting STOMP connection <0.19254.0> (XX:49217 -> YY:61613)
>
> =INFO REPORT==== 26-Sep-2013::16:12:46 ===
> closing STOMP connection <0.19254.0> (XX:49217 -> YY:61613)
>
>
> On Thu, Sep 26, 2013 at 4:07 PM, Simon MacMullen <simon at rabbitmq.com
> <mailto:simon at rabbitmq.com>> wrote:
>
> Hmm. Something must be establishing short-lived stomp
> connections though, do you see lines like:
>
> =INFO REPORT==== 26-Sep-2013::15:06:39 ===
> accepting STOMP connection <0.376.0> ([::1]:56780 -> [::1]:61613)
>
> in the log?
>
> Cheers, Simon
>
>
> On 26/09/2013 2:07PM, Santi Nuñez-Cacho wrote:
>
> It was the connetion to the mcollective rabbitmq middleware.
> ( we 've
> got all the machines connected to mcollective )
>
> I removed the mcollective conection. Disabling it. And
> restarting the
> rabbitmq-server.
>
> That's we got now
>
>
> [root at rabbit03.bcn rabbitmq]# lsof -i :61613
> COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
> beam.smp 17472 rabbitmq 21u IPv6 4652854 0t0 TCP
> *:61613 (LISTEN)
> [root at rabbit03.bcn rabbitmq]# rabbitmqctl eval
> 'length(supervisor:which___children(rabbit_stomp_client___sup_sup)).'
> 50
> ...done.
> [root at rabbit03.bcn rabbitmq]#
>
> And the process growing...
>
>
>
>
> On Thu, Sep 26, 2013 at 2:57 PM, Simon MacMullen
> <simon at rabbitmq.com <mailto:simon at rabbitmq.com>
> <mailto:simon at rabbitmq.com <mailto:simon at rabbitmq.com>>> wrote:
>
> On 26/09/2013 1:51PM, Santi Nuñez-Cacho wrote:
>
> I expect 0 stomp connetions. The unique one
> connection on 61613
> should
> be the cluster one.
>
>
> Port 61613 is not used for clustering, it's just stomp.
>
>
> [root at rabbit03.bcn ~]# rabbitmqctl eval
>
> 'length(supervisor:which_____children(rabbit_stomp_client_____sup_sup)).'
>
> 3802
> ...done.
> [root at rabbit03.bcn ~]# lsof -i :61613
> COMMAND PID USER FD TYPE DEVICE
> SIZE/OFF NODE NAME
> ruby 5638 root 6u IPv4 25097
> 0t0 TCP
> rabbit03.bcn.softonic.lan:____59851->rabbit:61613
> (ESTABLISHED)
>
> beam.smp 27183 rabbitmq 21u IPv6 4521334
> 0t0 TCP
> *:61613 (LISTEN)
>
>
> Ah.
>
> So you have one connection, from some ruby script. But 3802
> connection processes. So something is somehow leaking.
> Thank you for
> reporting this bug.
>
> But to help get to the bottom of it - what is that ruby
> script doing?
>
>
> Cheers, Simon
>
> --
> Simon MacMullen
> RabbitMQ, Pivotal
>
>
>
> --
> Simon MacMullen
> RabbitMQ, Pivotal
>
>
>
--
Simon MacMullen
RabbitMQ, Pivotal
More information about the rabbitmq-discuss
mailing list