Signaling Server - Connection Lost

[eunderkoffler]

We currently have two signaling servers, which load balance and one lost connection the other day for a couple of minutes. Not really sure why. Has anyone else seen this?

Thanks


[0000] 13/04/2010 08:44:19 LOG0004 tRUDPSS: Forwarding for connect Id = 414763108, error of type = 6 (connection lost).
[0000] 13/04/2010 08:44:19 LOG0004 tRUDPSS: Forwarding for connect Id = 414763108, error of type = 6 (connection lost).
[0001] 13/04/2010 08:44:19 LOG0003 tRUDPSS: ITS2008 Terminal connection status: 144.99.202.156 lost (20)
[0002] 13/04/2010 08:44:19 LOG0006 SET: 144.99.202.156 TN 97-17 Terminal offline
[0003] 13/04/2010 08:44:19 LOG0005 VTM: 144.99.202.156 Unregistered, terminal = 0x18ac6498, device = 0x18c1b4e8
[0004] 13/04/2010 08:44:19 LOG0004 tRUDPSS: Forwarding for connect Id = 414778468, error of type = 6 (connection lost).
[0005] 13/04/2010 08:44:19 LOG0003 tRUDPSS: ITS2008 Terminal connection status: 144.99.202.50 lost (20)
[0006] 13/04/2010 08:44:19 LOG0006 SET: 144.99.202.50 TN 98-17 Terminal offline
[0007] 13/04/2010 08:44:19 LOG0005 VTM: 144.99.202.50 Unregistered, terminal = 0x18ac5818, device = 0x18c17a28
[0008] 13/04/2010 08:44:20 LOG0004 tRUDPSS: Forwarding for connect Id = 414769764, error of type = 6 (connection lost).
[0009] 13/04/2010 08:44:20 LOG0003 tRUDPSS: ITS2008 Terminal connection status: 144.99.202.59 lost (20)
[0010] 13/04/2010 08:44:20 LOG0006 SET: 144.99.202.59 TN 80-16 Terminal offline
[0011] 13/04/2010 08:44:20 LOG0005 VTM: 144.99.202.59 Unregistered, terminal = 0x18ac6b78, device = 0x18c0f0a8
[0012] 13/04/2010 08:44:23 LOG0004 tRUDPSS: Forwarding for connect Id = 414767204, error of type = 6 (connection lost).
[0013] 13/04/2010 08:44:23 LOG0003 tRUDPSS: ITS2008 Terminal connection status: 144.99.202.72 lost (20)
[0014] 13/04/2010 08:44:23 LOG0006 SET: 144.99.202.72 TN 98-21 Terminal offline
[0015] 13/04/2010 08:44:23 LOG0005 VTM: 144.99.202.72 Unregistered, terminal = 0x18ac6c18, device = 0x18c19e68
[0016] 13/04/2010 08:44:24 LOG0004 tRUDPSS: Forwarding for connect Id = 414766436, error of type = 6 (connection lost).
[0017] 13/04/2010 08:44:24 LOG0003 tRUDPSS: ITS2008 Terminal connection status: 144.99.202.118 lost (20)
[0018] 13/04/2010 08:44:24 LOG0006 SET: 144.99.202.118 TN 97-19 Terminal offline
[0019] 13/04/2010 08:44:24 LOG0005 VTM: 144.99.202.118 Unregistered, terminal = 0x18ac5638, device = 0x18c19be8
[0020] 13/04/2010 08:44:29 LOG0004 tRUDPSS: Forwarding for connect Id = 414780772, error of type = 6 (connection lost).
[0021] 13/04/2010 08:44:29 LOG0003 tRUDPSS: ITS2008 Terminal connection status: 144.99.202.153 lost (20)
[0022] 13/04/2010 08:44:29 LOG0006 SET: 144.99.202.153 TN 97-04 Terminal offline
[0023] 13/04/2010 08:44:29 LOG0005 VTM: 144.99.202.153 Unregistered, terminal = 0x18ac5f98, device = 0x18c18ce8
[0024] 13/04/2010 08:44:32 LOG0004 tRUDPSS: Forwarding for connect Id = 414776932, error of type = 6 (connection lost).

 

[Tman45]

My guess is network issue Yeap looks like you lost half of your IP phones - that is a small bummer with load balancing - too bad Nortel doesn't offer a split-second failover on CS1000 - I think they do with OCS and SIP - cuz I thought during that demo they pulled a server connection and you barely saw the HD video stream blink Pretty cool.

 

[bignose21]

eunderkoffler need to look a bit earlier this is just a big list of the sig server timeout for registered sets. Look for where they started to drop and go a bit before. You cannot get the reason for failure from what you have posted just the effect.

Tman45 when you lose a sig server as above there should be no disruption to traffic either. A phone that has a call in progress will not do a full re-register to the other Sig Server until you hang up the call. All sets not in use will re-register on a timer, unless you are trying to use it to intiate a call then it will re-register almost straight away.

 

[Tman45]

Yep bignose21 - I didn't put it specifically like you did but based on eunderkoffler's post it looks like his phone TNs were dropping on the Sig Server that went down - so yes they were either hanging up their call or initiating a new call. I was just trying to point out to him that since his SigServers are load balanced that however many are on the server that went down they may re-register unless the downed server comes back before the timer reaches 0 - in his case it looks like he had a lot of phones re-reg.