Hi All,
I’m trying to work my way through the process to switch data centres in a DR scenario and then switchback following the recovery of the primary data centre. I don’t have a lab in which I can test this so it is all theory based…
Scenario
2 x data centres (AD DS Sites)
1 x Multi-Role Exchange 2010 Server (CAS, HT, Mailbox) in each data centre on Server 2008 R2 platform
1 x DAG Stretched across the 2 data centres
1 x CAS Array in each data centre
The primary data centre (Site A) hosts the Witness Server
The secondary data centre (Site B) hosts the Alternate Witness Server
Under normal operation Site A will be Active, Site B will be passive.
Namespace Site A (Primary Data Centre):
mail.x.x = internal namespace
mail.x.x/ OWA, ECP etc = Web Services URL
outlook.x.x = CAS Array
autodiscover.x.x
Namespace Site B (Secondary Data Centre):
mail2.x.x = internal namespace
mail2.x.x/ OWA, ECP etc = Web Services URL
outlook2.x.x = CAS Array
autodiscover.x.x
----------------------------------------
Switchover
1. If the primary data centre is online/accessible, there is network connectivity between the Data Centres and the Server in the primary site is online run:
Stop-DatabaseAvailabilityGroup –Identity <DAGNAME> –MailboxServer <SERVERNAME>
2. If primary data centre is NOT online/accessible, there is NO network connectivity between the Data Centres or the Server in the primary site is NOT online run:
Stop-DatabaseAvailabilityGroup –Identity <DAGNAME> –ActiveDirectorySite “Site A” –ConfigurationOnly:$True
My understanding is that the above commands will ensure that the primary data centre is marked as stopped by Active Manger – is this correct?
3. Stop the Cluster Services on the server in the Secondary Data Centre:
Stop-Service ClusSvc
4. Activate the Mailbox Server in the Secondary Site:
Restore-DatabaseAvailabilityGroup –Identity <DAGNAME> –ActiveDirectorySite “Site A” –AlternateWitnessDirectory:<ASWPath> -AlternateWitness Server:<ServerName>
5. Mount mailbox database and check status:
Get-MailboxServer | Set-MailboxServer -DatabaseCopyAutoActivationPolicy Unrestricted
6. Update DNS Records accordingly
Switchback
1. Force cluster cleanup in the primary data centre:
cluster Node <SERVERNAME.X.X> /forcecleanup
Ensure that the cluster service is DISABLED on the server in the primary data centre
2. Reset DAG
Start-DatabaseAvailabilityGroup –identity <DAGName> –ActiveDirectorySite “Site A”
3. Reset the File Share Witness:
Set-DatabaseAvailabilityGroup –Identity DAG1 –WitnessServer <SeverName> -AlternateWitnessServer <ServerName>
4. Dismount Mailbox Database in secondary site:
Get-MailboxDatabase | Dismount-Database
5. Mount Database in primary data centre:
cd “E:\Exchange Server\Scripts”
.\RedistributeActiveActiveDatabases.ps1 –DagName <DAGName> –BalanceDBsByActivationPreference
Get-MailboxServer | Set-MailboxServer -DatabaseCopyAutoActivationPolicy IntraSiteOnly
----------------------------------------
Do these steps look correct to those who have actually completed these tasks in the real world?
Your advice would be appreciated.
Cheers,
Andrew