I've set up an Enterprise OCS 2007 R2 environment which currently consists of a single front-end server. The machine was added to an existing AD infrastructure, all of which works flawlessly.
I've installed Office Communicator 2007 R2 on two laptop clients and provisioned two new test user accounts in AD which are OCS-enabled. All OCS configuration tests pass with flying colors, nothing in the event logs, etc etc etc.
Now here's where things get interesting...
This network uses split-brain DNS. For the sake of discussion, let's say that the domain names are internal.local (AD domain) and external.com. User accounts are therefore of the format:
joeuser@internal.local. Each user currently has three communications addresses assigned to him/her: an internal SMTP email address (e.g.
joeuser@internal.local), an external SMTP email address (e.g.
joeuser@external.com) and a SIP address (e.g. sip:joeuser@external.com). The external SMTP address is marked as default.
For the time being I am limiting my OCS testing to internal LAN only. As such, I have configured my two test laptops to sign into Communicator as follows:
Sign-in address: fred@external.com
User name: fred@internal.local
Password: ********
Sign-in address: barney@external.com
User name: barney@internal.local
Password: ********
On both clients I have chosen to manually specify the address of the pool, e.g. ocspool.internal.local, to eliminate possible problems with automatic signin.
Both of these clients log in just fine with no errors or warnings. Their presence indicator (big one at the top of the window) shows online. However, if I browse for Barney in Fred's client and add Barney to Fred's contact list, Barney's presence is shown as "presence unknown", and vice versa. More interesting, if Fred browses for Fred and adds himself to his own contact list, his presence is shown as "presence unknown" in his contact list, even though his status shown at the top of his window is Online. (Ditto for Barney of course.)
Trying to start an IM conversation with the other party, as you might have guessed, fails because the other party cannot be found.
Some digging at the logs show the reason quite clearly. Despite Fred's signin as
fred@external.com, when he browses for Barney and adds him to his contacts, he picks up
barney@internal.local, NOT
barney@external.com. For that matter, when Fred browses for Fred, he finds himself as
fred@internal.local. The resulting conversations therefore take the form of sip:fred@external.com trying to reach sip:barney@internal.local, and the latter address does not exist in AD, so the communication fails.
How can I resolve this? Is there a way to convince Communicator to look up SIP addresses instead of AD login names? Is there some way to convince it to return both? Or am I just doing something incredibly stupid or missing some obvious configuration step?
I suspect I could get all of this to work by changing our SIP naming standard to be sip:joeuser@internal.local instead of sip:joeuser@external.com, but IMHO that's a lousy solution, as sip:joeuser@exernal.com is clearly the more desireable and easily remembered form.
Thanks in advance for any light you can shed on this.
