IP Phones not coming up when switching over to Avaya ESS

[rtsoi]

I switched over to our ESS and everything came back up except for our 9621G IP Phones. We also have a few 1603's, which I didn't check but they probably had the same problem.

This was a forced switch so the actual gateways and everything was still up and running. I read elsewhere that because the G650s were still active, the phones were registering incorrectly.

Any help would be appreciated.

 

[kyle555]

ESS/LSP stuff isn't particularly complicated, but it's based on a certain sequence of events happening. If your test didn't replicate how that would happen with a WAN failure, then its very possible you didn't get what you were looking for.

Did any IP phones failover and back the way you expected and just the 9621s didn't?

Your ESS goes live when a gateway asks it for service. That means a G650's IPSI lost connectivity to the main server and knew to look for your ESS, or a smaller gateway lost connectivity with a CLAN or the main CM server and in its MGC list started looking for the ESS. One of those things has to happen before the ESS will let phones register to it. A CM server regardless of type - main/lsp/ess/whatever must have DSPs on it (so, a G350/450/etc or G650 with medpro) before it will let phones register.

When the connectivity failure happens, the phones will begin looking for the ESS based on the 'alternate gatekeeper list' defined in their network region.
Their network region is based on the ip-network-map where you define certain subnets as belonging to certain network regions.
In the absence of defining a subnet as belonging to a region, the phone will inherit the region of the ip-interface of the gatekeeper it registered to. So, either the network region of a CLAN or of procr/the main CM server.

Suppose you had no network map, and procr/CM was in network region 1 and network region 1 had no alternate gatekeeper list. That would be a great reason why your phones never looked for the ESS.

Suppose remote branch 12 had phones with gateway 12 and your dial plan was set to have all phones at branch 12 use the trunks in gateway 12 and you just went in gateway 12 and repointed it to the survivable server. So long as the phones in region 12 could reach the main server and so long as the main server had DSPs, the phones would never fail over, but their dialing would tell them to pick a trunk on gateway 12 no longer on their system and you'd have dialtone and a nice intercom system, but noone could make calls.

There are a few mechanisms you should be familiar with to test that stuff out. In newer CM releases, you can disable registrations within a network region. To say, you can issue a command in the main CM to "fail region 12!" and make the core CM not respond to that region thereby forcing the gateway and phones over to whatever provides failover service to region 12. There's also a deep little setting about "forcing phones and gateways to active ESS/LSP" to help mitigate split brain scenarios. To say, if for some silly reason, a G450 with LSP could no longer reach the main CM yet all phones at the site with that G450 could reach the main CM, when that G450 goes into survivable mode, the main CM would know to kick phones in the same network region off of the core so they failover with the gateway at the same site as they're at.

The section "reliability and recovery" of the IP Telephony Deployment Guide (

https://downloads.avaya.com/css/P8/documents/100016254)

outlines exactly how this stuff works.