Hi John,
Resuming here the discussion from LinkedIn group, There are 2 possibilities:
1) upgrade to latest stable 1.7.0 and hopefully the leak was fixed.
2) keep the version you have and enable memory debugging (see
http://www.opensips.org/Resources/DocsTsMem), set memlog=6 and memdump=1
(in this order) and when a process reports a mem outage, send to it a
SIGUSR1.
Regards,
Bogdan
On 10/18/2011 01:13 PM, John Quick wrote:
Using OpenSIPS version 1.6.2-notls with various modules enabled including
snmpstats.
The startup command line has the memory option set to "-m 256"
It is a busy production server handling several thousand calls a day. It
runs 24x7.
After opensips is running for about 1 week, I start to see errors and
warnings like the following:
Oct 4 08:11:36 SIP01 /sbin/opensips[17153]:
ERROR:snmpstats:insertContactRecord: no more pkg memory
Oct 4 08:11:36 SIP01 /sbin/opensips[17153]:
ERROR:snmpstats:executeInterprocessBufferCmd: openserSIPRegUserTable was
unable to alloc...
Oct 4 08:11:36 SIP01 /sbin/opensips[17153]: WARNING:core:fm_malloc: Not
enough free memory, will atempt defragmentation
OpenSIPS continues to process SIP messages and I can see no errors in the
call handling, but it concerns me that these errors are happening.
Could it be relevant that db_mode for usrloc is set to 3 - this means the
SNMP monitor always shows zero registrations.
John Quick
P.S. I posted this question before on 5th October but got no replies.
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
--
Bogdan-Andrei Iancu
OpenSIPS eBootcamp - 19th of September 2011
OpenSIPS solutions and "know-how"
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
