Hi Brett, The only way to hunt this down is to enable the mem debug (see http://www.opensips.org/Resources/DocsTsMem). Once compiled in, use: mem_log=10 mem_dump=1 debug = 2
This will prevent the mem runtime logs, but it will allow you to do the mem dump . After that, at runtime, when proc X reports the mem outage, simply do a "kill -SIGUSR1 X" to force proc X do dump the mem....and send me the dump to check. Regards, Bogdan Brett Nemeroff wrote: > On Fri, Jul 2, 2010 at 9:45 PM, <[email protected] > <mailto:[email protected]>> wrote: > > Hi Brett, > > While this doesn't address the root problem, we do two thing: (a) > run with ~100MB of pkg mem (per process), and (b) run a monitor > script checking syslog for out of memory messages, and restart all > the processes. Not elegant, but better than zombies. > > I noticed in your log that you're also out of shmem. Do you know > which comes first? (running out of pkg or shmem)? > > Regards, > Kennard > > > Kennard, > It appears that it runs out of pkg first, then a minute later or so > shm shows up too.. Weird thing is that I had converted this config > from mediaproxy to rtpproxy. With mediaproxy, we never had these > issues.. I'm not sure if that is actually related or not. > > Any ideas? > -Brett > > ------------------------------------------------------------------------ > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > -- Bogdan-Andrei Iancu OpenSIPS Bootcamp 20 - 24 September 2010, Frankfurt, Germany www.voice-system.ro _______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
