Configuring vSphere Replication#
VMware vSphere® Replication™ is integrated into the vSphere Web Client, allowing you to configure and manage replication for each VM.
To configure replication for a VM:
Select the VM you want to replicate.
Navigate to All vSphere Replication Actions.
Select Configure Replication.
This navigation brings up the Configure Replication wizard. Use the following settings for each section:
Replication type: Replicate to a vCenter Server®.
Target site: The correct location, typically your target VMware Cloud™ on AWS disaster recovery (DR) environment.
Replication server: Auto-assign initially only load balances VMs to servers and doesn't proactively move VMs to new replication servers. You can choose to manually select the replication server if you have more than one.
Target location: Select Edit to choose the target datastore for the VM. You can do this for individual VMs or many VMs at once. VMware recommends having twice the available storage at the target side of the VM you are replicating. If you're using vSAN™ on the target side, you need to account for the protection level (Failures to Tolerate (FTT) and Failure Tolerance Method) of the storage policy you are applying to the virtual machine disk (VMDK) at the target side. For example, a VM consuming 100 GB, an FTT of 1 requires 400 GB free raw space on the vSAN datastore (100 GB x 2 for vSphere Replication is 200 GB, FTT1 consumes twice as much on the vSAN datastore, thus 400 GB). vSAN displays available raw space without regard to FTT or Failure Tolerance Method because it is applied to the VM level and can differ depending on the storage policy. Use the following information to determine consumption on vSAN:
- FTT1 with Failure Tolerance Method of RAID-1 consumes 2 times the storage provisioned.
- FTT2 with Failure Tolerance Method of RAID-1 consumes 3 times the storage provisioned (five-node cluster minimum required).
- FTT3 with Failure Tolerance Method of RAID-1 consumes 4 times the storage provisioned (seven-node cluster minimum required).
- FTT1 with Failure Tolerance Method of RAID-5/6 consumes 1.3 times the storage provisioned (four-node cluster minimum required).
- FTT2 with Failure Tolerance Method of RAID-5/6 consumes 1.5 times the storage provisioned (six-node cluster minimum required).
If you're using VMware Virtual Machine File System (VMFS) storage, calculate consumption based on two times the storage provisioned for the VM.
This storage requirement is applied to both source and target VMs and datastores because the source site might become the target site in the event of a disaster.
- Guest OS quiescing (Optional) - This option stuns or pauses the VM during a replication process to halt disk modifications to keep the replica consistent. It might impact running services within the guest, leverage Microsoft®VSS for supported Windows® OS, and require scripts for non-Windows OS.
- Network compression (Optional) - This option might reduce throughput requirements for replication, but might also impact the recovery point objective (RPO) because there is higher CPU utilization to compress or decompress the data being replicated.
Recovery point objective (RPO) - The slider doesn't guarantee RPO and generates a vCenter alert if RPO extends beyond the configured time. Rackspace doesn't recommend going below 15 minutes.
Point-in-time instances (Not supported in Managed DR support level) - You can configure the number of snapshots to retain during replication.
These instances show up as snapshots on recovered VMs at the target site.
This situation can use drastically more storage at the target site because it functions similar to a snapshot retaining deltas.
Snapshot structure on the source-side VM is not maintained and is consolidated into the multiple point-in-time snapshots on the target side VM upon recovery.
Monitor vSphere Replication
- Select the home icon in the vSphere Web Client.
- Select Hosts and Clusters.
- Select the vCenter at the top of the left column.
- In the middle pane, select Monitor.
- Select vSphere Replication.
- Select either Outgoing Replications, Incoming Replications, or Reports for more information.
Point-in-time (PIT) instances (not supported in Managed DR)
- PIT retains replication instances as snapshots post-recovery, providing the ability to roll back to previous points in time of a recovered VM.
- Rackspace doesn't recommend enabling PIT instances because this might have significant storage requirements and has the potential to fill up the datastore at the recovery (target) site.
- If you choose to enable PIT instances, you should have a minimum of 4 times the amount (4x) of replicated data in free space at the recovery (target) site.