When the Production Site fails, the failover procedure of replicated VMs at DR Site with vSphere Replication can be initiated to quickly restore network services.
The failover procedure is an action that is performed quite easily in VMware vSphere Replication and you just need few steps to fire up replicated VMs.
Blog series
vSphere Replication 8.x deployment - pt.1
vSphere Replication 8.x failover - pt.2
vSphere Replication 8.x traffic isolation - pt.3
vSphere Replication 8.x seeding - pt.4
Perform failover
To check the status of replicated VMs ready to be powered on in case of need, access the Site Recovery and go to Replications area.
If the production virtual machine fails for any reason, its replica at DR Site can be fired up to quickly restore services provided.
To perform a failover, access the Site Recovery Replication page and right click the VM to be powered on. Select Recover from available options. Keep in mind that data stored in the replicated VM are updated based on latest replication occurred.
Select the Recovery option to use:
- Synchronize recent changes - available if the Production Site is up, this option should be used for planned maintenance or outage of the Production Site.
- Use latest available Data – in case of Production Site failure, this is the option to use.
You can also flag the Power on the virtual machine after recovery to power on VMs automatically. Click Next.
Specify the Folder for the VM at the Recovery Site then click Next.
Specify the Resource to use at Recovery Site then click Next.
Click Finish to power on the replica. Remember that network devices will be disconnected in the recovered VM to avoid possible conflicts with Production Site.
After a few seconds the selected VM is up and running in the specified DR Site with the Status reported as Recovered.
Looking at the vSphere Client you can see the recovered virtual machine.
To connect the recovered VM to the network, you need to connect the Network device. Right click the VM and select Edit Settings. Enable the Connect option for the VM's network adapter.
If the RPO policy has been configured, the recovered VM will have a certain number of snapshots available based on the RPO set. Selecting one of these snapshots you can revert the VM to a specific point in time.
To failback to production retaining changes occured in the VM, you need to configure a reverse replication from the DR Site to the Production Site.
If vSphere Replication is configured with default network, the appliance uses the management network to transfer replication data with the result of consuming the production network and affecting application performance. To avoid this situation, replication traffic should be isolated using a dedicated network. Part 3 will explain how to isolate the traffic to improve overall performance without using the management network to transfer the data.
