none
3 Node DAG File Share Witness Query

    Question

  • I have a 3 node DAG across 2 AD sites.  2 nodes are in one AD site and the other is in a separate AD site.  As its a 3 node DAG the FSW is configured for normal and alternate but is not used.  If I lose the primary site I fail over and come online using the alternate site FSW.

    When I fail back to the production site the alternate FSW stays as online in cluster manager for the DR site - the properties of the DAG show as no WitnessShareInUse though.  Before I fail over this is not here in cluster manager.  Is there a procedure to remove this?  Do I need to?  Do I need to worry about anything here, is this by design?

    Thanks.

    Monday, October 07, 2013 10:33 AM

Answers

  • Try it with no switches, yes. 

    I assume that all 3 nodes are up and running and happy.

    What is the FSW set to now? ( Not the alternate. the regular one)


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.


    Monday, October 07, 2013 3:20 PM
  • As soon as you run it, it asks for identity which I thought was the actual DAG name.  Didn't want to do any damage etc.

    Ah. Yes, enter the identity, i.e. the name of the DAG. It wont do any damage.


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.

    Tuesday, October 08, 2013 4:45 PM

All replies

  • What matters is what Exchange thinks. Once everything is back in order, ensure it shows as node majority for the DAG and you will be fine. You can run set-databaseavailibility wit no switches to "Clean" things up as well.


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.

    Monday, October 07, 2013 12:40 PM
  • Before the fail over the cluster was showing as Node majority.  Using the alternate file share witness and then failing back its now showing as Node and File share majority still using the alternate file share witness instead of the normal witness server.  Running get-databasavailabilitygroup -status shows WitnessShareInUse as Alternate.  Is it okay to leave like this, I guess it should be none currently - can I set this value to None without causing any damage or run the above like you say with no switches?
    Monday, October 07, 2013 2:41 PM
  • Try it with no switches, yes. 

    I assume that all 3 nodes are up and running and happy.

    What is the FSW set to now? ( Not the alternate. the regular one)


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.


    Monday, October 07, 2013 3:20 PM
  • running it with no switches seems to want to set it up again.  when running

    get-databaseavailabilitygroup NAME -status | fl

    it lists the witness server and alternate witness server fine.  WitnessShareInUse it lists as alternate.  Failover cluster manager lists this as well.

    Tuesday, October 08, 2013 12:18 PM
  • Not sure what "running it with no switches seems to want to set it up again" means.

    Running that command with no switches simply "sets" the config the way its already defined and often cleans things up.


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.


    Tuesday, October 08, 2013 12:20 PM
  • As soon as you run it, it asks for identity which I thought was the actual DAG name.  Didn't want to do any damage etc.
    Tuesday, October 08, 2013 2:56 PM
  • As soon as you run it, it asks for identity which I thought was the actual DAG name.  Didn't want to do any damage etc.

    Ah. Yes, enter the identity, i.e. the name of the DAG. It wont do any damage.


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.

    Tuesday, October 08, 2013 4:45 PM
  • Excellent - thanks.  Running that has fixed it and put it back how it should be.

    Brilliant :-)

    Thursday, October 10, 2013 8:43 AM