VMware replication feature allows to create an identical copy of a source VM on a target ESXi host in a powered off state ready to be powered on in case of a disaster.
Because of its powered off state, the copy of the VM (replica) doesn’t consume any resource and it is available in VMware vSphere. Up to 30 recovery points can be created for a VM replica and they are available as a regular VM snapshot allowing the administrators to return back to desired good state even without using Nakivo Backup & Replication.
With replication you can protect the most critical systems storing the copies of the VMs locally or offsite. If the source VM fails you can immediately failover from the source VM to the VM replica reducing dramatically the recovery time.
Compared to the classic restore from backups, replication allows to minimize the length of downtime causing only a minimal interruption of the network services.
Based on VM snapshots technology, the first replica is an identical copy of the source VM and supports sparse, thin, thick, and flat VM disk types. Using the VMware CBT technology, the next replication job is forever incremental since it merges only changed data blocks of the source VM since the last checkpoint creating a new recovery point. This process allows to increase dramatically the replication speed.
Configure a Replication job
From the Nakivo Dashboard, click Create menu and select VMware vSphere replication job.
Select the VM to replicate and click Next.
Specify the Container, Datastore and Network for the replica. Click Next to continue.
Set the schedule for the replication then click Next.
Replication can be set to run every minute to obtain a near real time replication.
Set the Job name, the retention and the requested parameters. Click Finish to save the job configuration. Using the App-aware mode option as Enabled, the product ensures that the data in applications and databases is always consistent.
A nice option is Replica disks that allows you to create thin disks on target VMs to save space.
Screenshot verification allows to verify the replica integrity to guarantee a successful VM recovery.
From the Dashboard, right click the Replication job just created and select Run Job to execute the job immediately.
Click Run to proceed.
The replication is executed.
When the job completes successfully, in the right pane you have all the job related info.
When the VM at the primary site fails, you can move the workload (failover) from production site to a DR site to quickly restore network services. The recovered VM from replica will contain all changes since last replication.
From the Dashboard, click Recover menu and select VMs from replica option.
Select the replica to recover and the desired recovery point. Click Next.
Specify the Network to be used and click Next.
Enter the Job name and click Finish to save the configuration.
Right click the just create job and select Run Job.
Click Run Recovery to power on the replica.
The Recovery job is executed.
When the job completes, in the right pane are displayed all the Job Info.
From the vSphere Web Client you can verify the running replica.
When the failed VM at the primary site is restored, you perform the failback process to return workloads to the primary site.
When you failback to production, two scenarios have to be considered:
- Discard changes – changes in the recovered VM can be discarded
- Commit changes – changes occurred in the recovered VM from replica have to be committed and moved to production
To discard changed data, from vSphere Web Client right click the recovered VM and select Power> Shut Down Guest OS option.
From Nakivo Dashboard, right click the previously created Recovery job and select Delete.
Select the option Delete job and recovered VMs then click Delete to confirm. All changes occurred in the VM are discarded.
Power on the source VM in the primary site to restore the service.
If you need to move all changes back to the production VM, you can replicate the recovered VM to production. From Nakivo Dashboard, right click the previously created Recovery job and select Delete.
Select the option Delete job and keep recovered VMs then click Delete to confirm.
Create a new replication job for the VM that you want to permanently recover and use the Use existing VM as a target option on STEP 2. This will transfer all changes to the VM in production.
To ensure zero data loss, power off the recovered VM in VMware vSphere client and run the VM replication job one again.
Power on the VM created by the replication job.
The VM in the primary site is now up and running with all changes occurred during the failover time.