Methods to Instantly Restore Virtual Machines

By CIOReview | Monday, September 12, 2016

Storage through virtualization such as Virtual Machines (VMs) has revolutionized the way IT operations are performed, diminishing the need to build large storage infrastructures across large physical spaces. As long as storage is concerned, disasters are one of the important factors, which CIOs need to focus on while choosing a VM product. The product selected should be able to heal itself instantly during any outage or disaster failure, as this will help the customers to seamlessly work without being interrupted. To mitigate these failures, delay in restoration of primary storage is the biggest hurdle and this brings the need to redirect the users from primary to secondary storage.

Redirecting is made possible through a feature called as Instant Recovery. This feature enables a backup system by taking snapshots of existing VMs and running it temporarily on secondary storage. This redirect to the secondary storage will continue till the primary storage revives from its original state and is ready for operation. With Instant Recovery practice users are left unaware of the events that are happening on the VMs and this offers a seamless workflow. Instant Recovery in virtualized environments is also known as ‘Recovery-in-place’ which has the ability to develop a data volume that can serve as backup during disasters. Along with these abilities a secondary storage should have the performance and efficiency similar to that of primary storage.

While leveraging data backup techniques, a CIO should look for data deduplication, backup frequency, and the user’s performance needs. Although having a disk backup is a cost effective feature, data deduplication has a performance laggings when restoring. There are some disk appliances which create their own non-deduplication landing zone while there are others which provide processors to compensate loss in performance. Also when executing a restoration, backup files are the most important ingredient. Backup file can be produced by continued push and pull of data every second. A backed-up file in the past will have no value as they lack latest updates or information which has been updated recently. So a backup frequency has to be setup which will help in getting last updated data during restoration. A CIO should confirm the frequency and ensure that enterprise employees follow it as a Standard Operating Process (SOP). Also, a CIO should check the level of performance that each of the clients or user requires when using secondary storage.

Instant Recovery has been deployed on VMware—Hyper-V by Veeam Software Company that develops backup, disaster recovery and virtualization management software for VMware Hyper-V. Veeam provides three steps for instant recovery. The first is initialization phase where Veeam Backup and Replication function collects the resources, which are necessary for restoration. It also interacts with the data repository for data transfer process. After initializing Veeam Data Mover, the second step is Network File System (NFS) mapping where Veeam vPower NFS Service is used to configure VM files and links to virtual disks that run as a backup on the repository. The final step is to restart the VM. After the VM failure, VMware vSphere runs VMs on Veeam NFS data store, disguising Veeam NFS data store as a regular data store. And with this, users can be redirected from the primary storage system to the backup storage system, enabling users to work seamlessly. 

As Veeam’s offer software for Hyper-V virtualized environments, the VM configuration is read from the backup and restores system, creating an empty disk for Instant Recovery. The next step is to take the snapshot of the VM in case of any disaster. The destination host is intelligent enough to identify the permanently stored data thus decreasing the workload of the restoration process. This improves the performance of the recovered VM. Also Hyper-V supports bulk restore of multiple VMs and uses the resource schedule mechanism to allocate and use the resources required for instant VM recovery.