thanks for the time. I had some questions regarding how DAGs and the activation preference works.
we have a 4 node DAG. 2 in one datacenter and 2 in another. but they are all part of the same DAG.
DC1:
-MBX1-DB1-active - preference 1,2,3,4
-MBX2-DB2-active - preference 2,1,3,4
DC2:
-MBX3- passive copy
-MBX4 - passive copy
Each DC has its own CAS array.
Now to the questions.
Besides the activation preference is there another way to keep the active copies in DC1...so when MBX1DB1 fails over it fails over to MBX2. I know that removing a member from the DAG is not a viable option since it remove the DAG copy. If the activation preference is the only option can I only choose to put only the servers from DC1 as option 1 and 2 and not put a preference on the servers in DC2? Also what is the algorithm as to why MBX2 is not chosen but fails over MBX3? Does Exchange think that MBX2 is busy and cannot handle it?
Second one...I get that when one of the database fails over to MBX3 or MBX4 that Outlook users will get prompted:
'The Microsoft Exchange administrator has made a change that requires you quit and restart Outlook'
this is because it is going from one cas array to another. However this shouldnt have if the database is failing from MBX1 to MBX2. correct?
And lastly how often should one see DAGs failover between nodes? I see this in the logs.
'A failover or switchover occurred'
Thanks for the time and input. sorry for all the questions. Im hoping to get some clarification on how DAGs work.