We are moving one of our datacenters to a new facility. In order to complete the move, we will need to completely shutdown our active messaging system, which is hosted in that site. I'm trying to verify that I have all the steps needs for successful shutdown and also to bring everything back online. I've searched around and I'm seeing some mixed feedback on all the steps required to do this.
Primary site has 2 mailbox servers, 2 HUBCAS servers, 2 UM servers, 1 TMG server. We have 2 databases and this site has HA for the mailbox servers with active and passive copies of each other's database. One of the HUBCAS servers is the primary File Share Witness.
DR site has one mailbox server, one HUBCAS server, one UM server, one TMG server. This site is DR only and has only passive copies of the databases in the primary site. However, we do not want to failover to the DR site for this event.
The steps I'm planning to follow are:
1. Shutdown MB servers by doing the following:
(This script should move all active DB to other DAG member, pause the node in cluster, suspend replication, set DatabaseCopyAutoActivation
Policy to block, and cluster core resource moved to other server in DAG)
Power off the server
ps1 -DR Mailbox Server
(This should prevent the DR server at the other site from becoming the PAM, activating databases, etc.)
Is this step necessary since this server is in our DR site?
Here is where some additional confusion comes in:
Shutdown the last remaining mailbox server at the primary site
Before I shut down the final mailbox server, do I need to run the StartDagServerMaintenance.
ps1 -Final Mailbox Server
script? Since it is hosting the active databases, PAM, etc. I'm assuming it will fail if I run that script. I've also seen some recommendations to dismount the databases on the active server at this point. Other recommendations are to just shut the server down without running any commands.
2. Shutdown disk enclosure for both mailbox servers at primary site
3. Shutdown HUBCAS servers
Shutdown the one serving as the FSW last
Do I also need to shutdown the HUBCAS server in the DR site since it is the Alternate FSW
4. Shutdown UM servers
5. Shutdown TMG server
6. Shutdown domain controller in the site that is relocating
Move the datacenter.
When bringing everything back online follow in this order:
Power on Domain controller in the site, then:
1. Power on TMG server
2. Power on HUBCAS servers
Make sure FSW is back online before powering MBX servers back on
3. Power on UM servers
4. Power on disk enclosures for mailbox servers
5. Power on mailbox servers
s1 script for the last mailbox server to be shutdown prior to move
Verify health of the databases
s1 for second mailbox server at primary site and redistribute the databases accordingly
s1 for DR mailbox server at the secondary site
6. Verify that replication has resumed for the database copies on all servers
Test-ReplicationHealth and Get-MailboxDatabaseCopySta
tus to verify the DAG health
7. Test all Exchange related services:
Test using microsoft remote connectivity anlayer
Test using microsoft remote connectivity anlayer autodiscover
Test using your script http://exchangeserverpro.com/health-...mailbox-server
Send inbound / outbound emails to verify routing
Also test basic stuff like opening Outlook and booking a meeting with a meeting room, or checking free/busy information for a meeting invitee.
Check the event logs on the servers for any unusual errors
If anyone is able to validate the steps I have listed and also make recommendations on anything I may have overlooked, as well as answer the questions above, I would greatly appreciate it.