I just changed my DNS domain on presence and it broke the HA failover.
I just did this via the CLI using set network domain. Looking at the HA state the secondary shows initializing and the primary shows "peer down during initialization" Checking into the server recovery manager logs from the secondary I see the following:
2011-11-28 10:47:19,877 ERROR [HL] rsrm.RsrmHeartBeatHandler - Rejecting message. Sender: LG-SPES-CUP01.lseg.stockex.local does not match peer host name: LG-SPES-CUP01.stockexchange.com or Ip addr: null
Setting the domain back to what it was before (stockexchange.com) fixes it and gets everything up and running. However, I need to change the domain.
Is there a specific process to do this? I've been unable to find anything relevant online, except in the Cisco docs it states something I've already done.
From the CUPS deployment guide:
Changing the Domain Value
Follow this procedure if you want to change the domain value (from one valid domain value to another valid IP proxy domain value).
This procedure is applicable if you have a DNS or non-DNS deployment.
Note It is highly recommended that you use a DNS deployment. In order to be considered valid, the domain value must match the DNS domain name. Using a valid domain name ensures that the SRM initializes correctly in a High Availability deployment.
Procedure
Step 1 Stop the Cisco UP SIP Proxy, Presence Engine and XCP Router services on Cisco Unified Presence on all nodes in your cluster.
Step 2 On the publisher node, perform the following steps to configure the new domain value:
a. Select Cisco Unified Presence Administration > System > Cluster Topology.
b. In the right pane, select Settings.
c. Configure the Domain Name value with the new domain.
a. Select Cisco Unified Presence Administration > System > Service Parameters, and select the Cisco UP SIP Proxy service.
b. Configure the Federation Routing CUP FQDN with the new domain.
c. You will be prompted to confirm these configuration changes. Select OK for both prompts, and then select Save.
Step 3 On all nodes in the cluster, use this CLI command to set the new domain:
set network domain <new_domain>
This CLI command invokes a reboot of the servers
The service parameters page stipulates:
This parameter specifies the public FQDN of the routing Cisco Unified Presence server when federating with Microsoft OCS/LCS. Note: This value MUST NOT be the actual FQDN of any Cisco Unified Presence server within the enterprise.
This makes me think its completely irrelevant to what i actually want to do as I'm not using OCS/LCS, however - its set to LG-SPES-CUP01-public.lseg.stockex.local anyway (not a valid FQDN; the host is junk; but the domain exists) and it still doesnt work once the domain is set to anything other than stockexchange.com