KubeVirt-Powered OpenShift: Managing VM Storage as Native Objects

Creating virtual machine snapshots in OpenShift

In a virtualization environment, snapshots play a crucial role by capturing the state of a VM’s disks at a specific time. This offers numerous benefits, such as rolling back to a known good state and restoring VMs in case of an issue. Furthermore, unlike cloning, one can take snapshots of a running VM.

Requirements

The following are required for VM snapshots in OpenShift virtualization:

• A supported storage provider: Any storage provider with a CSI driver that supports the Kubernetes volume snapshot API can be used to snapshot virtual machine disks in OpenShift virtualization.
• QEMU guest agent: While not strictly mandatory, installing the QEMU guest agent within the VM’s operating system is highly recommended for online snapshots. The guest agent facilitates communication with the storage driver, ensuring data consistency during the snapshot process. Snapshots might not capture the most recent data state without the agent, potentially leading to inconsistencies.

Creating snapshots

To check if your storage class provisioner supports taking snapshots, you can explore the documentation for the specific CSI driver. You can list the CSI driver for your cluster and then find the available snapshot classes:

$ oc get csidrivers
NAME                                 ATTACHREQUIRED   PODINFOONMOUNT   STORAGECAPACITY   TOKENREQUESTS   REQUIRESREPUBLISH   MODES     AGE
openshift-storage.rbd.csi.ceph.com   true          false         false                false            Persistent   115d
$ oc get volumesnapshotclasses
NAME                                     DRIVER                               DELETIONPOLICY   AGE
ocs-storagecluster-rbdplugin-snapclass   openshift-storage.rbd.csi.ceph.com   Retain        115d

To create an online snapshot for your virtual machine, create an offline snapshot for our VM:

$ virtctl stop my-fedora-1
VM my-fedora-1 was scheduled to stop
$ oc get vm
NAME       AGE STATUS READY
my-fedora-1   5m7s   Stopped   False

Define a VirtualMachineSnapshot object in your manifest file. The snapshot.kubevirt.io API group defines the necessary resources for snapshotting and restoring virtual machines. Specify the snapshot name and the source virtual machine:

apiVersion: snapshot.kubevirt.io/v1alpha1
kind: VirtualMachineSnapshot
metadata:
  name: my-fedora-1-snap
spec:
  source:
    apiGroup: kubevirt.io
    kind: VirtualMachine
    name: my-fedora-1

Create the snapshot from the manifest file, and verify that it has been completed and then make sure to check the status field in the output. In case of successful completion, the status will be shown as successful:

$ oc create -f snap.yaml
virtualmachinesnapshot.snapshot.kubevirt.io/my-fedora-1-snap created
$ oc get vmsnapshot
NAME            SOURCEKIND    SOURCENAME PHASE    READYTOUSE   CREATIONTIME   ERROR
my-fedora-1-snap   VirtualMachine   my-fedora-1   Succeeded   true
   3s 

To restore the VM from a snapshot, define a VirtualMachineRestore object in your manifest file. Specify the snapshot name and the source virtual machine:

apiVersion: snapshot.kubevirt.io/v1alpha1
kind: VirtualMachineRestore
metadata:
  name: my-fedora-1-restore
spec:
  target:
    apiGroup: kubevirt.io
    kind: VirtualMachine
    name: my-fedora-1
  virtualMachineSnapshotName: my-fedora-1-snap

The restore operation can only be performed on a powered-down VM, so make sure the VM is powered down:

$ oc get vm
NAME        AGE   STATUS   READY
my-fedora-1   6m4s   Stopped   False

Create the restore object and verify that the VM has been restored:

$ oc create -f restore.yaml
virtualmachinerestore.snapshot.kubevirt.io/my-fedora-1-restore created
$ oc get vmrestore
NAME                   TARGETKIND    TARGETNAME COMPLETE   RESTORETIME   ERROR
my-fedora-1-restore   VirtualMachine   my-fedora-1   true    3s 

Once the snapshot has been restored, the VM can be powered up again:

$ virtctl start my-fedora-1
VM my-fedora-1 was scheduled to start

Creating and restoring virtual machine backups in OpenShift

In addition to snapshots and clones, complete backups of virtual machines can be created in OpenShift. Unlike snapshots—which capture a point-in-time image—or clones—which generate a copy of a running VM—backups provide complete, restorable, independent copies of virtual machines for long-term retention.

Specialized tools can be used for backing up virtual machines. Trilio is one of the most popular solutions for creating backups in OpenShift and can protect both application containers and virtual machines. In the following demo, we’ll see how Trilio can be used to back up and restore virtual machines in OpenShift.

You can refer to this link to install and configure the Trilio operator in OpenShift. Once the Trilio operator has been installed, the Trilio Backups tab will become available in the OpenShift console.

Creating virtual machine backups using Trilio

To create virtual machine backups using Trilio, follow these steps:

1. Define a backup target: Specify the storage location for your backups.
2. Create a backup plan: Determine backup frequency, retention policy, and other relevant settings for your backup.
3. Execute backups: Initiate the backup process for your virtual machines based on the defined plan.

To create a backup target, go to Trilio Backups -> Target in the OpenShift console and define the storage location for your backups. The target storage can either be an NFS or an S3 object store. Here, we will use an NFS server to store our virtual machine backups. Make sure that you do the following:

• Select the desired namespace for the backup target.
• Assign the necessary write permissions on the NFS/S3 target to store the backups.
• Assign the necessary storage space to the backup targets so that they don’t run out of space when backing up virtual machine data.

Creating backup targets using Trilio

Once Trilio has validated your defined backup target, its status should change to Available.

Successful backup target creation

Now that we’ve created our backup target, we can store backups. At this point, we will switch to using the Trilio dashboard for the remaining steps. We can continue using the OpenShift console to complete the remainder of the tasks, but using the dedicated Trilio dashboard to manage backups is recommended. The Trilio dashboard is specifically designed for backup and recovery operations, offering a more comprehensive range of features, granular control, and advanced options than the OpenShift console. 

You can launch the Trilio dashboard as follows:

Launching the Trilio dashboard

Once you’re in the Trilio console, create a backup plan as follows. Select the “Backup & Recovery” tab  and click Create Backup and select “Application”.:

Creating virtual machine backups

Creating a backup plan

Select the virtual machine tab, select the virtual machines you want to backup and click “Next”.

Selecting virtual machine for backups

If there is no existing backup plan, create a new one by clicking “Create new”.

Creating a new backup plan

And then enter the configuration parameters as shown below. Select the target that was previously created for this backup plan.

Defining backup plan parameters

There are no additional components or applications to add so you can proceed to click next on the Component Details tab. We are not creating a Continuous Restore policy for this backup, so we can proceed to click on “Skip and Create”.

Successful creation of the backup plan

Once the backup plan has been completed successfully, click finish and you will return to the “Select Backup Plan” window.

Ensure that the newly created plan is selected and click “Next”. 

Selecting the newly created backup plan

Enter a name for the backup and click “Create Backup” to start the backup process.

The backup will take some time, depending on the size of your virtual machine. The dialog box will continue to show you the different stages of the backup until it is complete.

Once the backup plan is created, a success screen appears with a status log as shown below:

Successful backup creation

Restoring virtual machine backups using Trilio

The backup virtual machine can be restored in the same project or a different one. To restore a virtual machine from backup, select the View Backup & Restore Summary option from the Create Backup drop-down menu:

Restoring virtual machines from backup

Click the Restore option at the bottom:

Restoring from available restore points

Select the desired project where you want to restore the backup and set a name for the restore resource. You can skip the Advanced section and click Create. Once done, you would get a status log with a confirmation of the restore:

Successful restore operation

Once the restored operation completes, the restored virtual machine will run in the selected project.

Limitations of snapshots and clones in OpenShift Virtualization

Consider the following limitations when using snapshots and clones in OpenShift Virtualization:

• Not all storage classes in OpenShift Virtualization natively support snapshots or cloning. The underlying storage provider must support the OpenShift Volume Snapshot API.
• The QEMU guest agent must be installed in the operating system to create snapshots of live VMs.
• Although you can create snapshots for running VMs, you must power off the VM to restore a snapshot.

Optimizing Day 2 storage management in OpenShift

The following are some of the best practices for managing virtual machine storage in OpenShift virtualization:

• When creating PVCs for VMs, request only the necessary storage capacity to avoid wasting space.
• Estimate the disk usage of your VMs and factor in snapshot sizes to ensure sufficient storage capacity for both active VMs and snapshots.
• Choose an appropriate storage provider that supports snapshots and clones.
• Regularly review and delete old snapshots that are no longer needed. Retaining unnecessary snapshots can consume valuable storage space.
• Schedule snapshots during periods of low VM write activity to minimize the risk of data inconsistencies.
• Use a comprehensive backup and disaster recovery solution for your virtual machines to ensure data protection and quick recovery in case of an incident. Consider using solutions such as Trilio for this purpose.