I was able to work with Byron Spurlock (former Microsoft Consultant with MCS) to have him deliver another Webcast Series on Lync Server 2010.  The second one was delivered on September 23, 2011 on Lync Server 2010 High Availability.  I am a bit behind, but I am now catching up and posting the Questions and Answer Log from that session here.  You can also find the PowerPoint deck on my SlideShare site here: http://www.slideshare.net/harold.wong/lync-2010-highavailability.

If you would like to get a hold of Byron, please feel free to contact him at Byrons@Quadrantechnologies.com.

Questions and Answers Log: TechNet Webcast - Microsoft Lync 2010 High Availability and Resiliency Deep Dive


Question: What is a Registrar Pool? it is not one of the Lync server roles, so what exactly is a Registrar?
Answer: When the Lync client signs in, it will register with the Front End Server or Director. The Director is running the registrar service just like the Front End Server.


Question: any known issues around draining/ I tried and it disconnected my current sessions
Answer: I have not seen issues where turning on draining caused current sessions to be disconnected. That is contrary to what the feature is supposed to do.


Question: So the HLB will round robin through the 3 front ends and the Front Ends will then talk to each to figure out what the primary registrar is for a client..correct? If so..what happens if the the primary registrar goes down..will a new registrar be chosen for the client?
Answer: In a resiliency configuration, the user is assigned a backup registrar pool when they register with their primary registrar.


Question: Does Lync support SIP DNS load balancing with SRV records, instead of multiple A records?
Answer: No, you need to have multiple A records.


Question: in dns lb why aren't cnames used for ocspool1 hosts?
Answer: You are technically creating an A record for each server in the pool that happens to have the same name as the pool. The CNAME record points to an A record so you have to have the A record first.


Question: What are the latency requirements between the DC to DC failover options and SBA to DC failover scenario?
Answer: There are not any specific “latency” requirements for failover to another DC or from SBA to DC. As a right provisioning standard, provision your network to ensure a maximum end-to-end delay (latency) of 150 milliseconds (ms) under peak load.


Question: the a records are created by building the front end servers so a records for these IPs are already created so then when you configure your pool and dns lb you can create cnames for the pool to the server names.... this way if the server IP changes the pool still stays alive. right?
Answer: The Pool needs to have an A record to function.


Question: When we need to use a Director?
Answer: The Director is not required, but recommended depending on your configuration.


Question: So Full services are available for remote user who's home server (SBA/SBS) is down,,what about if user is homed to another pool? Do they have limited functions?
Answer: Provided the resiliency is configured correctly for all pools, then they would have full functionality.


Harold Wong