Please refer to the following Simulation Result for 2 x Exchange 2013 DAG Nodes Fail-Over & Fallback within a Single AD Sites for one of our client
- 1 x Exchange 2013 Server installed on each Building and being connected using 1Gbps Fiber Connection
- File Share Witness (FSW) Server installed in each site
Pre-Fail-Over Stage
- Exchange Mailbox Database are mounted in AVENTIS-EX01 and the replication is running without error for VIP & General Mailbox Databasees
*Notes for File Share Witness (FSW) Configuration **
- Added **Exchange Trusted Subsystem **as Local Administrators Group in both Exchange 2013 Servers
- Create a Share Folder (FSW), and share it by assigning FULL permission to Exchange Trusted Subsystem
- Enable Data Center Activation Mode (DAC Mode) to prevent the Split Brain (Split brain is prevented using DAC mode, because DAC mode requires DAG members to obtain permission to mount databases before they can be mounted)
Set-DatabaseAvailabilityGroup -Identity DAG -DatacenterActivationMode DagOnly
Scenario #1 – Fail-Over
Move All Active Mailbox Databases to AVENTIS-EX02, and shutdown AVENTIS-EX02 & File Share Witness (FSW)
Both VIP & General Mailbox Database are show DisconnectedAndHealthy in AVENTIS-EX01
Stop Clustering Service in AVENTIS-EX01 Server
#Stop the Clustering Services
Get-Service | ? Name -like "ClusSvc" | Stop-Service
Remove the Failed Node from DAG Cluster
#Remove the Failed Node from DAG Cluster
Stop-DatabaseAvailabilityGroup -Identity DAG -MailboxServer AVENTIS-EX02 –ConfigurationOnly
Restore DAG Cluster by pointing to Alternative FSW
#Restore the DAG by pointing to Alternative FSW
Restore-DatabaseAvailabilityGroup -Identity DAG -AlternateWitnessServer AVENTIS-RDS08 -AlternateWitnessDirectory C:\fsw
All Mailbox Databases are online and mounted successfully in AVENTIS-EX01
Additional Information There is only 1 node in DAG Cluster now
AVENTIS-EX02 is in StoppedMailboxServers
Scenario #2 – Fallback
Power On AVENTIS-EX02, and noticed the following
- Cluster Services is Stopped
- All Exchange Services are running
Join the Failed node back to DAG, and start seeding from Active Database Copy
#Join the Failed Node back to DAG, and start seending from Active Database Copy
Start-DatabaseAvailabilityGroup -Identity DAG -ActiveDirectorySite Default-First-Site-Name
You can safely ignore the warning message and verify that all mailbox database are back to normal
Optional Steps
You can move / point the FSW back to the original or just leave the FWS stay where is it
Set-DatabaseAvailabilityGroup -Identity DAG | Set-DatabaseAvailabilityGroup -WitnessServer VCENTER
That all for the Exchange 2013 Fail-Over & Fallback simulation in our lab and hope it helps.