[ejabberd] Server allocating all available ram, then stops at 100% CPU

Evgeny Khramtsov xramtsov at gmail.com
Tue Nov 12 14:13:50 MSK 2013


В Mon, 11 Nov 2013 23:04:25 +0100
Peter Schwindt <ejabberd at schwindt-net.de>:

> Evgeny Khramtsov schrieb:
> 
> > Do you have other {process_limit, {max_queue, N}} crashes with N
> > less than 38992?
> 
> Like these?
> 
>     exception exit: {process_limit,{max_queue,1493}}
>      Reason:     {process_limit,{max_queue,1493}}
>     exception exit: {process_limit,{max_queue,1326}}
>      Reason:     {process_limit,{max_queue,1326}}
>     exception exit: {process_limit,{max_queue,1291}}
>      Reason:     {process_limit,{max_queue,1291}}
>     exception exit: {process_limit,{max_queue,1157}}
>      Reason:     {process_limit,{max_queue,1157}}
>     exception exit: {process_limit,{max_queue,1056}}
>      Reason:     {process_limit,{max_queue,1056}}
>     exception exit: {process_limit,{max_queue,1095}}
>      Reason:     {process_limit,{max_queue,1095}}
>     exception exit: {process_limit,{max_queue,26845}}
>      Reason:     {process_limit,{max_queue,26845}}
>     exception exit: {process_limit,{max_queue,1024}}
>      Reason:     {process_limit,{max_queue,1024}}
> 
> Happened right before the server stopped working.

Yes, like this. This means that max_fsm_queue is working actually.
So the problem is in "blocking" TCP send. What Erlang version are you
using?


More information about the ejabberd mailing list