[rabbitmq-discuss] Rabbitmq Crash and number of consumers

Fri Oct 7 09:08:54 BST 2011

Hi all,

We are using Rabbitmq 2.5 version and yesterday we found (a number of time) in the log file this entry:

=INFO REPORT==== 6-Oct-2011::14:58:13 ===
    alarm_handler: {set,{file_descriptor_limit,[]}}

Finally it hit an error:

=ERROR REPORT==== 6-Oct-2011::14:58:13 ===
** Generic server msg_store_persistent terminating
** Last message in was {'$gen_cast',
                           {client_dying,
                               <<199,5,182,196,129,67,147,182,61,201,143,0,
                                 51,197,190,234>>}}
** When Server state == {msstate,
                            "/var/lib/rabbitmq/mnesia/nodename/msg_store_persistent",
                            rabbit_msg_store_ets_index,
                            {state,520256,
                                "/var/lib/rabbitmq/mnesia/nodename/msg_store_persistent"},
                            22,#Ref<0.0.0.1817>,
                            {dict,0,16,16,8,80,48,
                                {[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],
                                 []},
                                {{[],[],[],[],[],[],[],[],[],[],[],[],[],[],
                                  [],[]}}},
                            [],undefined,589,5530342,[],<0.296.0>,524353,
                            516159,528450,
                            {set,19,16,16,8,80,48,
                                {[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],
                        ....
** Reason for termination ==
** {{badmatch,false},
    [{rabbit_msg_store_ets_index,insert,2},
     {rabbit_msg_store,write_message,3},
     {rabbit_msg_store,handle_cast,2},
     {gen_server2,handle_msg,2},
     {proc_lib,wake_up,3}]}

=INFO REPORT==== 6-Oct-2011::14:58:14 ===
stopped TCP Listener on [::]:5672

Right after this, rabbitmq closed the listener socket and did not recover. The service kept running without accepting any connections, to fix the problem we needed to restart the service.
What can be happened?

I have got another question: when I run the rabbitmqctl list_queues consumers the number of the consumers how is it calculate?
There is a "live" calculation or the information are stored somewhere and refreshed after "n" seconds?
I've got a queue with a single consumer and sometimes (I don't know why!) the consumer disappears, to detect that there is a script running which uses rabbitmqctl list_queues command but It can detect this situation with a delay of about 10-15 minutes even if the script  runs every minute.

Best Regards
--
Andrea Rosa

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20111007/cc49608b/attachment.htm>