Use the VMware vStorage API for Data Protection to backup virtual machines

Learn how VMware's vStorage API for Data Protection enables backup of virtual machines.

Implementing data protection for server virtualisation environments can be tricky. Conventional, agent-based, file-level techniques for virtual machine backup could cause problems -- most notably resource contention, which could compromise the performance of all virtual machines (VMs) running on an ESX or ESXi host.

Data backup vendors, aware of this pitfall, advocate the use of VMware Consolidated Backup (VCB) (now renamed vStorage API for Data Protection in vSphere 4) for virtual machine-level backup. VMware vSphere offloads backup processing from the VMware server, eliminating the backup window. However, VCB/vStorage API for Data Protection is a framework for backup -- not a backup application.

VMware vStorage API for Data Protection facilitates LAN-free backup of virtual machines from a central proxy server rather than directly from an ESX Server. It provides a set of drivers and scripts that enable conventional backup applications to perform file- or virtual machine-level backup of running virtual machines. In either case, the backup application instructs VCB/vStorage API for Data Protection to obtain a snapshot of the virtual machine at the prescribed time. The snapshot is made available to the proxy server (where a backup application client agent is resident) by mounting the snapshot (file-level backup) or exporting the image (VM-level backup); the data is backed up by the backup application, and VCB/vStorage API for Data Protection unmounts the virtual machine snapshot and takes the virtual machine out of snapshot mode. Backup policies and the central catalog are maintained by the conventional backup application.

VMware Data Recovery

With the introduction of VMware vSphere 4, its next-generation virtual infrastructure, VMware introduced a backup "utility" called VMware Data Recovery (VDR). VDR is an agentless, disk-based (any VMFS storage, including direct-attached storage, iSCSI, NFS or Fibre Channel storage, or CIFS shares) backup product that employs snapshot and data deduplication. Users of vCenter Server now have built-in backup capabilities for up to 100 VMs per appliance (with one appliance per vCenter license).VMware Data Recovery is included in vSphere Advanced, Enterprise and Enterprise Plus editions. It can also be purchased a la carte with vSphere Standard edition.

VMware Data Recovery runs in a virtual machine as a virtual appliance. Data backup and recovery tasks are managed via vCenter Server, which automatically discovers virtual machines. Through a workflow wizard, backup policies and schedules are set to automate processes. vCenter monitors virtual machines and backup jobs -- even when a virtual machine is moved by VMware High Availability (HA), VMotion, and DRS -- to ensure that backups continue.

VMware Data Recovery actually uses VMware's backup framework (the aforementioned vStorage API for Data Protection) to perform rapid VM-level backup and recovery. However, file-level recovery is also possible from the VM snapshot. Because VDR takes advantage of the vStorage API for Data Protection, incremental snapshots (a new feature of the API) is available. Combined with block-level data deduplication, disk space is kept to a minimum.

The drawback of VDR? VMware Data Recovery is only enabled for operational recovery. It has no way to move data offsite for disaster recovery purposes -- i.e., no ability to replicate data, no facility to create portable media, such as tape, and no integration capabilities with third-party tape-based backup.

More "built-for-VMware" backup: PHD Virtual, Veeam and Vizioncore

VMware Data Recovery is intended for virtual -- not physical -- server environments. There are vendors, such as PHD Virtual Technologies, Veeam Software and Vizioncore Inc. that only focus on backup and recovery of virtual environments. Of these vendors, Vizioncore is the only vendor with a backup solution for ESXi.

PHD Virtual has a VMware-only backup solution, esXpress, which is similar to VMware Data Recovery. It runs as a virtual appliance that can be managed from a browser or Virtual Center and is licensed per ESX host. The Linux-based virtual machine is only running during the backup schedule (starts up and before backup and hibernates after). This approach means that you don't need client agents in virtual machines or the service console, and it doesn't require VCB (or the associated proxy hardware or scripting). Because it's a virtual machine itself, it is VMware HA-, VMotion- and DRS-aware/compatible. One or more Virtual Backup Appliances (or VBAs) performs daily backups of all virtual machines on the hosts within the defined backup window. It supports VMFS so it can write to any data store -- FTP, CIFS, VMFS volumes -- but not tape. esXpress has source-side global deduplication capabilities across an ESX farm. For recovery, you can restore at the file- or VM-level. Its autonomous data restoration provides full VM recovery without requiring esXpress or other VM infrastructure.

Arkeia Software  Network Backup, EMC Avamar, and FalconStor Continuous Data Protector also run as virtual appliances and have many of the same qualities as the other backup virtual appliances mentioned. However, these producs also provide data protection for physical environments, as well as support hypervisors other than VMware.

It should come as no surprise that VMware added basic backup capabilities to its platform; Microsoft has had a basic backup utility for years. For SMBs that don't need much customisation or sophistication in backup, VMware Data Recovery will fit the bill. Environments with broader requirements have plenty of options with VMware-specific backup vendors, as well as the abundant number of data backup providers whose roots in physical server backup have spread to the virtual world.

Next Steps

Use APIs to create a custom disaster recovery report

Read more on Virtualisation and storage