Exchange failback - options?

[jwalker]

Hi all

Apologies if this is in the wrong section, or if it's been asked before, I had a quick search and could not find anything, so here it goes:

The power in my Production site is going to be down for 24 hours, and I have the option of running some services from my DR site, including Exchange. The Exchange environment is 4 servers in a HA cluster (2x mailbox server + 2x CAS/Hub server).

I am planning on mounting all mailbox DBs on the first mailbox server, then shut down the second mailbox server and start the second mailbox's replica in the DR site. Once the replica is up I plan to mount the mailbox DBs on that, which will unmount them on first mailbox server. Then I'll shutdown and failover the first CAS/Hub server to DR, which will give the DR site all the resources it needs to service email during my Production site's power outage.

Once the power outage is over, I'll start up my Production site, starting the second CAS/Hub server and then start the first mailbox server. Once the first mailbox server has its act together, all services are running, the Mailbox Database Copy has successfully synced etc, I will mount the mailbox DBs on the first mailbox server.

Assuming mail is now successfully being served via the first mailbox server + second CAS/Server in the Production site, how should I deal with the replicas in DR?

Is it best to simply shutdown the replicas, and restart their counterparts in the Production site?

Or should I undo the failover for each, or do a failback for each?

Any thoughts would be appreciated, thanks for reading!

[Vitaliy S.]

Hi jwalker,

If you are going to transfer/sync all the changes from replicas to production Exchange VMs manually (via Exchange native functionality), then you can shutdown VM replicas/undo failover without doing actual failback operation.

Thanks!

[jwalker]

Hi

That makes sense, thanks. For the sake of consistency across the state of the entire VM I am still tempted to failback, but will do some testing around this.

Thanks again

[jwalker]

It worked like a charm.

[veremin]

Thank you for taking your time and confirming that the proposed solution works for you; much appreciated.