Using VM Replication
There are multiple scenarios in which VM Replication can be used. At a minimum, a single PHD VBA can be used in one virtual environment to replicate VMs from one host to another. This is done by creating backups for your VMs, then creating replication jobs that will create replicas of those VMs on another host in the same virtual environment.
Figure E – 1. Using a single PHD VBA to replicate VMs within the same virtual environment.
When using a DR site, you could deploy another PHD VBA to that site to run your replication jobs. In this scenario, the PHD VBA at the DR site would be configured to connect to the primary site’s backup storage directly, then use that data to create the replica VMs on the hosts at the DR site. Then as backups are created at the primary site each night, replicas can be updated with the changed data from each backup.
Figure E – 2. Replicating VMs from a primary site to a DR site.
Before replicating VMs, you must have at least one PHD Virtual Backup VBA deployed and running regular backups. These backups can be stored on any supported backup data storage (attached virtual disk, a CIFS share, or an NFS share). Note that if an attached virtual disk is used, it must be shared in order to be used for replication. For details, see Connectors tab
In addition, a stable network connection must exist between any primary and secondary locations in order to transfer replicated data.
After establishing your regular backups, you will create replication jobs to run after the backups have finished to replicate any changes.
Step 1: Deploy/Configure a PHD VBA for Replication
If you will be using a DR site, the first step to configuring replication requires that you deploy a new PHD VBA to your secondary location. This is the location where you will be replicating VMs to (the “DR site” in the image above). The PHD VBA deployed here will read the data from your primary site’s backup location then use that data to create replica VMs. Follow the instructions in the PHD Virtual Backup Installation Guide to deploy a new PHD VBA.
After deciding which PHD VBA will be used to perform the replication, it will need to be configured. Open the PHD Console and connect to the location where your replication PHD VBA is deployed. In the Configuration area, configure the hypervisor credentials (if it is a new PHD VBA) on the General tab (refer to the Installation Guide for additional detailed deployment and configuration instructions).
Next, click the Replication tab. This is where you will add backup storage locations to be used with replication.
Click Add to add a backup storage location. These can be backup storage locations used by another PHD VBA (at your primary site, for example) or the current PHD VBA’s backup storage. Add all of the backup storage locations from which you want to replicate VMs.
After entering the required information and clicking OK, the storage location is displayed in the list, as seen in the following image.
You can automatically discover any backup storage for all PHD VBAs in your environment, by clicking Discover. More information about the Replication tab is in the section Replication tab.
After adding the storage locations you will use for creating replica VMs, you will next use the Create Replication Job wizard to create your replication jobs. The wizard can be opened from the Jobs area of the PHD Console, or from the Replication area.
From the Replication area, in the Virtual Machines Available for Replication tab, you can see all of the VMs available for replication. This list of VMs represents all of the VM backups available on the storage locations you added in the previous step.
At this point, you can select the VMs you want to replicate then click Replicate to open the Create Replication Job wizard. Follow the steps in the wizard to create the replication job. For details, see Using the Create Replication Job Wizard.
Tip: Replication takes advantage of existing backups to create and update replica VMs. Therefore, some consideration should be given to when your replication jobs are scheduled to run. If possible, replication jobs should be set to run after your nightly backups complete, to ensure your replicas are up to date with the latest changes.
After the job is created, it is stored in the Jobs area of the PHD Console, as seen in the following image.
Step 3: Run Replication Jobs and Manage Replicas
When a Replication Job runs, either scheduled or if manually started, VMs are replicated to the location you specified. If you view the Replicated Virtual Machines tab in the Replication area of the PHD Console, you will see all of the VMs that have been created.
Note: Only one replica VM can be created per hypervisor environment (per vCenter or host). For example, if you create a replica of VM1 on host1 and VM1 is also included in a second replication job replicating to that same host, when the job runs, the original replica will be found and no new replica will be created. A warning stating the replica already exists will be logged and included in the email report when the job completes.
From the Replicated Virtual Machines tab, you can see the status of each replica VM and select them for either testing or failover. For additional information about testing replicas and failover, refer to the following sections.
- When the virtual disks of a replicated VM change, for example, if a disk is added or removed, or if a disk’s size is altered, the next replication job will include a warning and any changed disks will not be replicated. In order to include virtual disk changes to your VMs, you will need to remove the existing replica and then either edit or create a new replication job to make sure any disk changes are included. You can verify changes are included by viewing the VM in the View/Edit step of the Replication wizard.
- VM configuration settings are not updated after the initial replication takes place. For example, if you replicate a VM that was configured to use 1 GB of memory, the replica VM will be configured with the same value. If you then change the size of the memory on the source VM, the replica VM will still have the original 1 GB value. To update the replica, you can either manually change the settings on the replica VM, or delete the replica VM and let the replication job create a new full replica the next time the job runs. This also applies to the replica’s configuration settings – if you change a setting on the replica, it will not be overwritten by incremental changes applied during a replication job.
- The list of VMs available for replication is automatically synchronized with the configured backup storage locations every five minutes. You can force synchronization by clicking refresh on the Replication page of the PHD Console.