Lync 2013 High Availability

When designing your Lync 2013 topology it is easy to overcomplicate the topology with Enterprise Pools in each datacenter.  Spanning pools is not supported across metropolitan datacenters in Lync 2013 which means as a minimum recommendation you would require 3 Front End servers in each datacenter  pool.  Then on top of that you require SQL backend servers, Edge Servers, Reverse Proxies, Office Web App Servers, File Servers, Load Balancers and Persistent Chat servers!  Not to mention you require Active Directory, a Certificate Authority and Exchange UM if you want Voicemail!

When designing my first Lync 2013 topology it started to get very big.  After some careful consideration I decided to implement 2 Standard Edition Front End servers as a “Pool Pair” across the Primary and Secondary datacenters.  This was due to the Topology being for less than 5000 users as well as the Lync servers being hosted in Tier 3 datacenters on a highly available VMware cluster, SAN storage and Network.  The chances of a FE server failure were low and a Standard Edition Pool Pair provides an adequate MTTR if that server was to fail (<30 minutes with Administrator intervention).

If one of the servers was to fail in the Pool Pair, users will failover to the other server automatically.  Users will receive an error in their Lync Client “Limited functionality due to an outage”  but will still be able to make phone calls and IM.  They will lose features like contacts, conferencing and Presence.  However with Administrator intervention you can manually failover the pool to the other member of the Pool Pair using the Invoke-CSPoolFailover command.  After this is complete, contacts, conferencing and presence are all restored.  Here is the Technet article on how to failover a pool:

Enterprise Voice calls will still work assuming your Voice Routing is configured correctly on your PBX/Gateway.  Response Groups also need to be recovered manually during a failover.

It is also worth pointing out that you can balance users across both servers in a Pool Pair (e.g. 2500 on one, and 2500 on the other) so the impact is less if a server was to fail.  In my environment the server in the secondary data centre is for failover (DR) only. I am currently writing a PowerShell script that will sit on the DR server that can be executed by a System Administrator should the primary Front End fail.  This can be used to failover all services to the secondary Front End in a DR scenario.  I will post it on here once complete and tested.

Finally, here is an excellent blog post that explains the advantages of Standard Edition servers in a Pool Pair to provide HA and DR:

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.