[rabbitmq-discuss] rabbit chokes?

[Owen Nelson]

Hi all,
I've been running rabbitmq for some web services (limited intranet user
base) for about 6 months now and everything has been great... until
recently.

I started to notice that the rabbitmq service would randomly die. 
Digging through the logs I started to see messages like the following:

=INFO REPORT==== 1-Nov-2009::09:13:41 ===
starting TCP connection <0.27459.0> from 127.0.0.1:47608

=INFO REPORT==== 1-Nov-2009::09:13:41 ===
closing TCP connection <0.27459.0> from 127.0.0.1:47608

=INFO REPORT==== 1-Nov-2009::09:13:43 ===
    alarm_handler: {set,{system_memory_high_watermark,[]}}

=INFO REPORT==== 1-Nov-2009::09:13:44 ===
    alarm_handler: {clear,system_memory_high_watermark}

=INFO REPORT==== 1-Nov-2009::09:13:51 ===
Rolling persister log to
"/var/lib/rabbitmq/mnesia/rabbit/rabbit_persister.LOG.previous"

=INFO REPORT==== 1-Nov-2009::09:14:16 ===
    alarm_handler: {set,{system_memory_high_watermark,[]}}

=INFO REPORT==== 1-Nov-2009::09:14:17 ===
accepted TCP connection on 0.0.0.0:5672 from 127.0.0.1:47609

=INFO REPORT==== 1-Nov-2009::09:14:17 ===
starting TCP connection <0.27501.0> from 127.0.0.1:47609

=INFO REPORT==== 1-Nov-2009::09:14:17 ===
closing TCP connection <0.27501.0> from 127.0.0.1:47609

=INFO REPORT==== 1-Nov-2009::09:14:18 ===
    alarm_handler: {clear,system_memory_high_watermark}

=INFO REPORT==== 1-Nov-2009::09:14:19 ===
accepted TCP connection on 0.0.0.0:5672 from 127.0.0.1:47612

=INFO REPORT==== 1-Nov-2009::09:14:19 ===
starting TCP connection <0.27508.0> from 127.0.0.1:47612

=INFO REPORT==== 1-Nov-2009::09:14:19 ===
closing TCP connection <0.27508.0> from 127.0.0.1:47612

=INFO REPORT==== 1-Nov-2009::09:14:19 ===
    alarm_handler: {set,{system_memory_high_watermark,[]}}

=INFO REPORT==== 1-Nov-2009::09:14:22 ===
    alarm_handler: {clear,system_memory_high_watermark}


The server that is running the web application, and the rabbitmq service
rarely feels any kind of strain, so I'm having a hard time believing
that it's suffering from resource starvation.  Any ideas as to what I
can do to not have to jump start rabbitmq every couple days?

Regards,
Owen Nelson