In a Lync 2010 ( as well as 2013) environment, I have a trusted application running as an automaton.
Now, we would like to provide a failover capability in the trusted app. Load balancing is not a requirement. We are not currently using either DNS load balancing or a hardware load balancer.
To this end, we created a trusted application pool with 2 computers, A and B, and are running a copy of the application on each. Each instance is configured to use its own FQDN, and uses a common certificate. Both instances connect fine as part of
the common pool. We have a SIP identity associated with the common pool.
How should we ensure a predictable failover scenario in this case? We would like to designate one instance as primary, and another as secondary.
At the moment, it appears that Lync routes the incoming sessions to the first to start application instance. For example, if application instance B started first and connected to the Lync Server, and then A started, incoming user sessions are routed to instance B. If B is shutdown, then sessions are directed to A.
What algorithm is used by Lync to decide where to route the sessions? Is there a predictable way to designate A as the primary, and B as the secondary.
This would also mean that if A recovers after a failure, Lync should start routing all new sessions to A and stop sending them to B.
Can this be done without using a hardware load balancer?
Thanks
(The reason for primary/secondary is that the application is stateful and we do not currently have a way to share information between the two instances of the application etc. In future we may do this using a shared DB or another shared memory technique but at the moment the application is not so enabled)
UCFin
