Instant VM Recovery is a new feature in PHD Virtual Backup. This feature provides a quick recovery from backup and by using this feature with PHD Motion you can recover back to a production ready state with minimal downtime if your not lucky enough to have Storage VMotion licensed with vSphere. I tried this feature out and it worked very well but I did have one minor issue while working with the beta version of PHD Virtual Backup 6.0.
So here is how Instant VM Recovery works. Using the Full VM backup method your vmdk files are backed up consistently as a whole. PHD Virtual has put together a way to utilize these vmdk files directly from the backup store. When your Instant VM Recovery job is kicked off, a virtual machine is created/restored on one of your production datastores, which you get to choose during the Instant VM Recovery wizard. Everything is restored but the vmdk files. The vmdk files are attached to the virtual machine from the backup store location. This allows for a much faster recovery on the front end because it doesn’t have copy the vmdk files from the backup store to a production datastore. During the normal restore process you’d also have to wait long periods of time until the restore was finished in order to power on and use the services/applications hosted on the virtual machine. This is almost similar to using PHD Virtual Backup with Replication where you could power on the replicated virtual machines for testing or an actual failover. The big difference is that everything is local in the primary backup store with Instant VM Recovery. PHD Motion completes the solution for Instant VM Recovery but giving customers a way to move back into a production datastore and out of the backup store.
Let’s get into the process a little bit more and add some pictures so you’ll have a reference of what this looks like in PHD Virtual Backup. First thing is to configure the Instant VM Recovery storage destination. You’ll get to either choose a dedicated storage location or one shared with the backup storage location.
Once the storage location is setup you can start the wizard for creating a Instant Recovery VM.
Choose the virtual backup appliance.
Select the virtual full backup to be used for the Instant Recovery VM.
Setup a few parameters for the job that will be used when creating the Instant Recovery VM. In the screenshot below I selected to append to the name of the original VM and power the VM on after it’s created. You have to also select the storage location for the virtual machine files and network port group.
The job gets created and starts the process of the creating the Instant Recovery VM with the virtual machine files in the production datastore but the vmdk data files will be run from the PHD backup storage location.
The status of the job is shown in the Instant VM Recovery location but you may have to manually refresh to get updated information.
Once the Instant Recovery VM is created you can see that the backup storage location is used for the data vmdk file.
You can also see that the virtual machines configuration, swap, etc files are created in the production datastore without the vmdk files.
To get the vmdk files into the production datastore PHD Motion comes to the rescue. Start PHD Motion from the Instant Recovery VM section.
The vmdk files are created/copied from the one being used in the backup store.
Once the PHD Motion is finished copying the vmdk file, one more step is required to switch from the PHD backup store to the one that gets PHD Motioned.
You will need to shutdown the virtual machine to complete the PHD Motion.
When the PHD Motion is completely done, the vmdk file is switched to the production file.