This document was generated from CDN thread
Created by: Ashraf Khandwalla on 13-04-2010 05:42:49 PM
hi,
We have a Top-Down based Route Group configured with two SIP trunks for Recording Servers. It takes approx 5 seconds for the second Recording Server to get a SIP INVITE if the first computer in the Route Group is powered off. This happens for all calls while the first computer is down and results in the first 5 seconds of every call not being recorded. We are able to adjust the duration by modifying the SIP TCP Timer in the Service Parameters. We are not sure if adjusting this timer is a good idea, considering that it is a system wide configuration and affects all trunks.
It looks like the Call Manager always tries the first trunk first regardless of whether the first trunk is up or not. Desired behavior would be that the Call Manager detects the failure on the first call; and send subsequent INVITEs to the second server until the first server was back up and running.
Is this behavior configurable or is there a recommended approach to handling this scenario so that we don¿t lose recording the first few seconds of a call?
Thanks.
Subject: RE: SIP trunk failover
Replied by: Giggesh Thekkekeloth on 28-06-2010 07:11:54 AM
It seems, in AngelFire 8.5 release, OPTIONS Ping feature is introduced to address the issue. For now there are two ways to work around. 1 as mentioned below, adjust 100trying timer. 2 use TCP instead of UDP. 2 will help the case of socket error. If CCM can create a socket to the remote recording server but somehow recording server just doesn't respond to the INVITE, 2 will not help. You still need to wait for time out.
