I've seen DAG migrations come down to two basic strategies:
- svMotion the DAG members as-is without making any application-level changes over to the new environment.
- Migrate the passive DAG member using svMotion to the new environment. Then failover the database from the active node (still on "old" env) over to the passive node (on "new" env).
In many cases, option #1 works just fine, but if your storage isn't particularly high performing or you have network bandwidth issues, it could cause a failover during the process. If that does happen, most of the time I've heard there was no application disruption, but still some prefer option #2 because it's predictable and within their control.
Thanks for the reply.
To be on the safe side I think I will go with option #2 and perform a server switchover, migrate the server to the new infrastructure, do some testing with a test DB and user on the migrated server and then repeat for the other 3 servers. I just wanted to mostly make sure that vMotion and svMotion with exchange was not inherently dangerous or capable of causing corruption on the server being migrated as I know exchange can be temperamental.
For a level of comfort, however a your mileage may vary warranty applies, we have ~6TB of Exchange 2010 mailboxes across 3 DAG members (6 DAG members in total). We have moved this from an HPE EVA (FC) to 3PAR 8400 (AFA) to a 3PAR 8200 (7.2K spinning rust) in two locations (primary and secondary sites) without issues.
In all these cases the servers never broke a sweat. When I did the move from the EVA to 3PAR 8400, I took option #2 as it appeared to me to be the safest route. Another engineer did the move from the 8400 to the 8200 without failing over the databases. No issues experienced with the databases or corruption.
Because of the enhancements svMotion (aka the mirror driver) there is a much greater level of confidence that the svMotion will complete in a timely and efficent manner however I prefer option #2 as any issues don't impact LIVE but that might just be me and you
Noted that you also mentioned about vMotion. The only changes we made to enable a trouble free vMotion experience were to modify the Cluser Heartbeat settings for the DAG. I'm sure you are also well aware of this document but in case you are not, I've linked here for reference: https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/solutions/vmware-microsoft-exchange-2013-best-practices-guide.pdf (Page 21 is where the settings are documented and recommendations made)
Thank you for the response. I am currently migrating one of four VM's now. Fingers are crossed!