Hello Everyone and thanks in advance for your expert Exchange 2013 tips, insights and advice.
We have two Exchange 2013 servers (running on 2012R2) with a DAG setup in the past by a vendor.
It had a database availability group witness server and witness directory for cluster failover and it worked.
However, I had to put in a new file server and needed to move the witness directory which I thought I did right, but apparently that is not the case.
Note that the Exchange Trusted Subsystem is a member of the local admin group on the new file server.
Some FYI information to help you help me :-)
The errors indicate (from what I can tell) that the witness folder is not available (permissions or something else), but it is created.
The share is there for sure and the share permission is everyone full control and the ntfs security permission includes all the right groups, servers, dag etc. as far as I can tell as I replicated the previous setup.
The failover cluster has not changed on the server itself and shows both exchange servers as online nodes
Command output on the Exchange shows cluster group online, with both Exch member servers part of DAG, shows the witness pointing to right server and share, but the quorumresource file share witness is listing tghfs1 which was the old file server... is that the issue and how do I change that...
Not good with power shell commands... and if I recall then I did run a command to point it to the new file server tghsvrfile1 and its share... Example: Set-DatabaseAvailabilityGr
oup -Identity DAG1 -AlternateWitnessDirectory
o.com -AlternateWitnessServer CAS3 but maybe it did not work or I got it wrong
I also ran some extra commands and the output seems to suggest that the file share witness does indeed point to tghfs1 (even though it is set to point to tghsvrfile1 in the GUI on the exchan admin)
Not sure what is wrong.
The witness server will not switch to new one until and unless there is a failover or 1 node goes down and only one node remains. Since both the nodes are up and running they form a cluster.
You do not need to do anything with regards to witness server, since if you have already changed to a new witness server this change has taken place but will not reflect and will reflect on its own when there is a failover or one nodes goes down.
Try to shutdown one server during non-business hrs. and see what cluster does with regards to having only one node up and new witness server.