I’ve been working on migrating an Exchange environment to 2010. This process includes an upgrade to the Unified Messaging role of Exchange to 2010 as well. We had purchased a UCC certificate to include all the Subject Alternate Names our Exchange environment would need and I had already applied it to the CAS server successfully. Since this certificate also included the FQDN of our UM server, I added the cert. and assigned the UM service to it so that Exchange could start processing voicemails through 2010 instead of 2007. [more]

After I had moved a couple of mailboxes (including mine) over for testing, I discovered that I could no longer receive voicemail. People were redirected to the Subscriber Attendant instead of my individual mailbox. There were events logged on the UM server saying the following:

Event ID: 1400 Source: MSExchange Unified Messaging
The following UM IP gateways did not respond as expected to a SIP OPTIONS request. Transport = TLS, Address = lync.ourdomain.com, Port = 5061, Response Code = 0, Message = This operation has timed out.

After considerable troubleshooting, I ran across a forum posting (http://social.msdn.microsoft.com/Forums/en-US/communicationsserversdk/thread/2733adf2-c91b-4b94-bf43-be93cdd2d5ef/) from someone who encountered the same issue and called for a support incident with Microsoft to figure out what was going on. In order for UM to work in a Lync phone system environment, the Subject Name of the certificate installed must be the FQDN of the UM server itself. It won’t work if it’s just included as a Subject Alternate Name; it must be the Subject Name.

I generated a new certificate from our internal CA with the UM server as the SN of the certificate, installed and assigned it to the UM roles (leaving our UCC cert running the remaining roles), and immediately started receiving voicemail notifications.

I found one other blog posting after the fact that backed this claim up even more (http://therealshrimp.blogspot.com/2012/03/multiple-exchange-um-servers-and.html).