locked
Communicator Timeout Values and failover RRS feed

  • Question

  • I am just running through some tests to see how long the communicator client takes to fail over to another front end server if the server it is on fails.  I am trying this on an R1 pool and have seen the client take up to 5 minutes to sign back in.  It takes even more time externally when accessing an edge server.  Are there values that can be specified to shorten this time?  Perhaps there are keep alive settings somewhere that we can set up?  Thanks.
    Wednesday, May 6, 2009 8:55 PM

All replies

  • No one has attempted to do this?  Or perhaps I have something misconfigured and the client should not take this long to re-register with the pool?
    Monday, May 11, 2009 12:46 AM
  • The client's don't really fail-over in the traditional sense.  When the Front-End service stops responsing, the client will disconnect but then repeatidly attmept to sign-back in to the same destination (via Automatic lookup or Manual configuration).  If using a director that is aware of the failed FE or changing SRV records to manually point to a diffrerent pool then you'll see different responses as DNS record changes might take a while for the client to pickup with manually flushing the cache.  I don't believe there is a configurable setting on the OC client to modify the time it waits before attempting to sign-in again.
    Jeff Schertz, PointBridge | MVP | MCITP: Enterprise Messaging | MCTS: OCS
    Monday, May 11, 2009 12:40 PM
    Moderator