Still not sure if it's a bug and should be added as a JIRA or if it is an expected behavior for that kind of load.

PS: I tried tweaking "max-db-handles" but it doesn't seem to make any difference with sqlite3 core db in tmpfs.

Regards,
François.

On Thu, 2014-12-18 at 12:21 +0100, François Delawarde wrote:
Hi all,

In a decent server (6x core E5-2420 with 16GB RAM) and under a bit of load, I see that FS sends NOTIFY up to a 5 minutes AFTER the actual event, so BLFs are useless when it happens.

Is it a normal/expected behavior? Code suggests that FS can queue up to 50000 events, and presence does lots and lots of DB stuff.

This particular site handles around 4000-5000 subscriptions and up to 2-3 calls per second under load. I use sqlite3 presence.db in tmpfs.

Is there anything that can be done to reduce that huge NOTIFY delay, or do I just have to live with it?

Thanks,
François. 
_________________________________________________________________________
Professional FreeSWITCH Consulting Services: 
consulting@freeswitch.org
http://www.freeswitchsolutions.com

Official FreeSWITCH Sites
http://www.freeswitch.org
http://confluence.freeswitch.org
http://www.cluecon.com

FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org