In two previous posts I talked about data protection of both vSAN and VVOLs volumes in VMware vSphere using Veeam Backup & Replication. Even if these two technologies are different in many aspects, one thing they have in common is they both leverage SPBM (Storage Policies Based Management) to control how virtual disks are placed and managed over these two storage solutions.
SPBM allows virtualization administrator to remove all the burden of manual placement of virtual disks, spreadsheets full of data about which VM is stored where, which LUN coming from a given array has feature X enabled, and so on. With SPBM, admins can create multiple policies with the needed options, and once the policy is applied to a VM, vSphere will automatically check for the compliancy of the VM and the storage it is actually stored onto, and if the policy is not fulfilled, a storage vmotion will happen to move the VM into a complaint storage. And policies can also be changed in real–time, and remediation again will happen automatically.
This new solution is a huge advantage, and many admins are leveraging this capability more and more. But what happens when a virtual machine has to be restored from a backup? Are those policies preserved? The answer is yes, if you use Veeam Backup & Replication.
SPBM policies over VVOLs
Let’s do a test. In my vSphere 6.0 environment I have a VVOL storage array (Nimble CS300) and I’ve created two storage policies like these:
Also, there’s the default VVOL policy created by vSphere itself, and another one created by one of my colleagues. By default, any virtual machine I deploy over VVOLs gets the VVol No Requirements Policy, unless I select one of the other two during the deployment phase, or I have it already specified in a template. Everytime I select a policy, I can check with compatible storage I have for that policy:
And I can also verify which VMs are stored using that policy:
SPBM policies backup and restore
Now, I want to protect my virtual machines by running a backup job. I create a Veeam backup job and I start to have multiple restore points of each VM. Note that in the job there’s no option to backup the SPBM policies, they are automatically saved during a backup everytime they are found, and in case the VM has a policy change, the new value is associated in the backup job together with the new restore point.
During a restore, Veeam gives me the possibility to restore the appropriate storage policy, or change it. By chosing to modify the VM parameters during the restore, Veeam drives the user through a series of options that can be changed. At the datastore step, the original volume and storage policy (in brackets) is shown:
The user can then select the entire VM, or even a single virtual volume and change the datastore. The popup lists the available datastores, datastore clustes, but also the different storage policies available:
In this way, upon restore I can see select different storage policies, or even restore a VVOL virtual machine on a VMFS or NFS datastore. The correct policy will be applied so that the VM will stay in place.
I believe like VMware that the future way to manage datacenters is by policies, and I like the fact that Veeam is already aligned to this vision.