[rabbitmq-discuss] Socket Error 104 when connecting to rabbit cluster

[Rubo Liang]

> I don't think it is a socket issue because I can connect successfully
with 'telnet'.

tcpdump gives more information(rabbit also binds to 10.4.16.12:5672):
1. the tcp connection is successfully setup
2. client sends a packet to server
3. server sends a packet to client
4. client sends another packet to server, server replies with ack followed
by a reset packet.

 tcpdump output:
18:38:08.396524 IP 192.168.20.41.61353 > 10.4.16.12.5672: Flags [S], seq
3696040838, win 14600, options [mss 1460,nop,nop,sackOK,nop,wscale 9],
length 0
18:38:08.396716 IP 10.4.16.12.5672 > 192.168.20.41.61353: Flags [S.], seq
2115356865, ack 3696040839, win 14600, options [mss
1460,nop,nop,sackOK,nop,wscale 7], length 0
18:38:08.396869 IP 192.168.20.41.61353 > 10.4.16.12.5672: Flags [.], ack
2115356866, win 29, length 0
18:38:08.397156 IP 192.168.20.41.61353 > 10.4.16.12.5672: Flags [P.], seq
3696040839:3696040847, ack 2115356866, win 29, length 8
18:38:08.397467 IP 10.4.16.12.5672 > 192.168.20.41.61353: Flags [.], ack
3696040847, win 115, length 0
18:38:08.397899 IP 10.4.16.12.5672 > 192.168.20.41.61353: Flags [P.], seq
2115356866:2115357201, ack 3696040847, win 115, length 335
18:38:08.397993 IP 192.168.20.41.61353 > 10.4.16.12.5672: Flags [.], ack
2115357201, win 31, length 0
18:38:08.399485 IP 192.168.20.41.61353 > 10.4.16.12.5672: Flags [P.], seq
3696040847:3696041097, ack 2115357201, win 31, length 250
18:38:08.436617 IP 10.4.16.12.5672 > 192.168.20.41.61353: Flags [.], ack
3696041097, win 123, length 0
18:38:11.401038 IP 10.4.16.12.5672 > 192.168.20.41.61353: Flags [R.], seq
2115357201, ack 3696041097, win 123, length 0


On Sun, Jul 14, 2013 at 6:17 PM, Rubo Liang <liangrubo at bytedance.com> wrote:

> Hi,
>
> I didn't take any action on this during this period of time and the
> problem remains( Server B and server C are clustered and I can connect to
> neither B nor C).
>
> I don't think it is a socket issue because I can connect successfully with
> 'telnet'.
>
> Logs in the log file seems abnormal:
> 1. no client is connected to the cluster, but rabbit emits "=INFO
> REPORT====" and "=WARNING REPORT===" pairs to the log file every 5
> seconds(see below).
> 2. 10.4.16.221 is a virtual ip which rabbit binds to, the log shows
> connections are from 10.4.16.221, why is this?
>
> log example:
>
> =INFO REPORT==== 14-Jul-2013::18:06:01 ===
> accepting AMQP connection <0.19551.11> (10.4.16.221:42919 ->
> 10.4.16.221:5672)
>
> =WARNING REPORT==== 14-Jul-2013::18:06:01 ===
> closing AMQP connection <0.19551.11> (10.4.16.221:42919 ->
> 10.4.16.221:5672):
> connection_closed_abruptly
>
> =INFO REPORT==== 14-Jul-2013::18:06:06 ===
> accepting AMQP connection <0.19556.11> (10.4.16.221:42920 ->
> 10.4.16.221:5672)
>
> =WARNING REPORT==== 14-Jul-2013::18:06:06 ===
> closing AMQP connection <0.19556.11> (10.4.16.221:42920 ->
> 10.4.16.221:5672):
> connection_closed_abruptly
>
> Thanks,
> Robert
>
>
>
>
> On Fri, Jul 12, 2013 at 7:08 PM, Michael Klishin <mklishin at gopivotal.com>wrote:
>
>> Rubo Liang:
>>
>> > I guess this won't cause the problem I got, right?
>> >
>> > I left the problem as it was for further investigation, is there any
>> suggestion on debugging this problem?
>>
>> correct, this likely is caused by your socket issue, not the other way
>> around.
>> --
>> MK
>>
>>
>> _______________________________________________
>> rabbitmq-discuss mailing list
>> rabbitmq-discuss at lists.rabbitmq.com
>> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>>
>>
>
>
>


-- 
梁汝波
手机：+86-18600464396
地址：北京市海淀区知春路甲48号盈都大厦B座10层
邮编：100088
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130714/0e6fc2b0/attachment.htm>