Anyone got experience with how limit scales with db back end?
I am using limit with db (sqlite) backend to limit all calls into my server. I have 72 inbound channels on gateways sending to my freeswitch, which is running on a single processor dual core 3GHz server, with the sqlite call limits database on the local disk, and at times I am seeing a delay of up to 3 seconds between the arrival of an invite and FS sending a 180 Ringing from the ring ready application. Other delays in signaling I also see intermittently. The only real processing going on before the ring_ready application is executed is the limit application. I see from the code there is a global mutex that looks like will result in all limit checks being single threaded through the code, and I am wondering if I am experience slight intermittent bottlenecks at peak call times. I realize the hash backend is faster and I may switch to that (decision pre-dated my arrival), but I am wondering if anyone out there has any good data about how many cps the limit application with the db backend can scale to. Obviously its dependent on hardware, but I'd still like to get any real world data (or suggestions) that are out there.
Is there any reason it's being stored on disk rather than ram? A waste of i/o of there's not.. -Avi Marcus On Tue, Nov 29, 2011 at 5:22 AM, Dave Horton <daveh@beachdognet.com> wrote:
I am using limit with db (sqlite) backend to limit all calls into my server. I have 72 inbound channels on gateways sending to my freeswitch, which is running on a single processor dual core 3GHz server, with the sqlite call limits database on the local disk, and at times I am seeing a delay of up to 3 seconds between the arrival of an invite and FS sending a 180 Ringing from the ring ready application. Other delays in signaling I also see intermittently. The only real processing going on before the ring_ready application is executed is the limit application. I see from the code there is a global mutex that looks like will result in all limit checks being single threaded through the code, and I am wondering if I am experience slight intermittent bottlenecks at peak call times. I realize the hash backend is faster and I may switch to that (decision pre-dated my arrival), but I am wondering if anyone out there has any good data about how many cps the limit application with the db backend can scale to. Obviously its dependent on hardware, but I'd still like to get any real world data (or suggestions) that are out there. _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org http://www.freeswitchsolutions.com
FreeSWITCH-powered IP PBX: The CudaTel Communication Server http://www.cudatel.com
Official FreeSWITCH Sites http://www.freeswitch.org http://wiki.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
participants (2)
-
Avi Marcus
-
Dave Horton