Had recently ran an upgrade from OCS 2007 R2 to Lync Server 2013 Technical Preview and stumble into a couple of ‘problems’. With these experience, I hope that it can be useful to those whom are undergoing the same activity as well:

  1. Make a FULL backup on the existing infrastructure – Active Directory, System State, SQL, OCS 2007 R2 and etc.
  2. Update the existing OCS 2007 R2 or Lync Server 2010 with the latest Cumulative Updates
  3. On the base operating system that would be installed with Lync Server 2013, make sure you install all updates as well
  4. If the OCS 2007 R2 is integrated with Voice Mail features from Exchange 2010 (I’ve not tested on Exchange 2007 so far), make sure the platform is on the latest Service Pack 2 with its Cumulative updates as well
  5. Next clean-up any settings that is not used/not optimized at OCS 2007 R2 before running the merge topology
  6. Once confirm that these changes has no impact to the enviroment, then only proceed with the migration/upgrade that you intended to.

Here’re a errors and resolutions during the process of upgrading:

Problem 1: Error Merging OCS 2007 R2 Topology

Under OCS 2007 R2 Console, enter the Global Configurations (Right Click Forest > Global Properties > General). Make sure the SIP domains that you’ve entered matches the ones at Lync Server 2010/2013. If not, remove them from OCS 2007 R2 or add those SIP domains into the Lync Server 2010/2013 via the topology builder.

Problem 2: Failure [0xC3EC79F2] Mediation server cannot be deactivated as it is being referenced by one or more phone routes.

Resolution: If you’ve imported/migrated all of the call routes previously, make sure that you’ve also a backup copy of the .RTDA file which can be exported using the Enterprise Route Helper. Delete all routes and proceed to deactivate the mediation Server and the process should be successful.

Problem 3: Failure [0xC3EC7A3E] There are still conference directories associated with this pool when attempting to remove pool. Please move the conferences before removing the pool.

And if you’ve used the Force option, you’ll get an error while publishing the updated copy via Topology Builder:

Error: An error occurred: “System.InvalidOperationException” “Cannot publish topology changes. Conferences still exist on one or more deleted services.”

Resolution: Delete the Conferencing Directory from the Lync Server 2013 PowerShell cmdlet, to do so, launch the Lync Management Shell and type the followings:

Locate the Conferencing Directory ID (in numeric number) that is hosted in the OCS 2007 R2 Pool Server:

> Get-CSConferenceDirectory

After identifying the number, execute the following command:

>Remove-CSConferenceDirectory -identity #

Problem 4: Deactivating & Clean-up the Front-End Pool in Sequence

Note: Saved the Front-end Server for the last!!

  1. A/V Conferencing Server
  2. Web Conferencing Server
  3. Web Components Server
  4. Application Sharing Server
  5. Response Group Service
  6. Conferencing Attendant
  7. Conferencing Announcement Service
  8. Outside Voice Control
  9. Application Host

If you decide to go against the ‘flow’ you’ll received an error message:

Failure [0xC3EC7A2E] There are still applications activated for this Office Communications Server 2007 R2, Application Host. Use LcsCmd.exe with /force switch to force task execution.

My recommendations: Please follow the sequence as suggested.

Problem 5: Unable to make remote PSTN calls due to dedicated Network for Voice (Mediation <-> Media Gateway)

To give a better clarity to the problem stated above, here’s my recent experience:

Site A

  • Subnet
  • Lync Servers and Media Gateways are residing at the same network subnet
  • Site-to-Site VPN with Site B, allowing 192.168.1.x/24 traffic to route to this local site

Site B

  • Network Subnet of 192.168.1.x/24
  • Lync Servers are residing on the 192.168.1.x network while the second Network Interface of the Mediation and Media Gateway is residing on a seperate network – 10.0.0.x/24
  • Site-to-Site VPN with Site A, allowing 192.168.0.x/24 traffic to route to the local site

After transiting to Lync Server 2013, both sites were only able to make local outbound PSTN but not cross sites. After investigating, I realized that the whenever Site B attempts to make a cross site outbound PSTN call, the network trace shows that the 10.0.0.x/24 segment wasn’t able to reach to the 192.168.0.x/24 network segment and eventually receives a request timeout. Resolution to this was Site B’s Media Gateway segment had to move into the same subnet as the rest of the Lync Servers which is 192.168.1.x/24 segment.

During the Office Communications Server 2007 R2 days, it is recommended to have the Voice Network (http://technet.microsoft.com/en-us/library/dd441140(v=office.13).aspx) and Mediation Servers are the first point of contact should a PSTN call is required. The importance of having this consideration in mind is if you’re infrastructure is spread across different counteries and each country has 1 Mediation and 1 Media Gateway. When switching to Lync Server 2013, you if intend to take advantage on Media Bypass you may want to move the Media Gateway to the same network subnet where the Lync Servers are placed, another main reason to do this is to ensure the Media Gateway network routing between different sites can be successful.

PS: I’ll keep building this list as my notes all went missing after upgrading to Office 2013 Technical Preview 😦