11-09-2010 12:33 PM
Companies have a number of virtualization vendors from which to choose – and many IT organizations are deploying two or more hypervisors to suit different needs. While VMware is widely acknowledged as the leader in virtualization, we see Microsoft Hyper-V appearing in more and more – especially in remote or branch offices where customers still want virtualization to provide high availability and save on hardware costs.
Tight integration between Microsoft Hyper-V and the Volume Shadow Copy Service (VSS) framework provides application data consistency in a Hyper-V environment. HP Data Protector software leverages this infrastructure to support a variety of methods for protecting Hyper-V virtual servers, all of which can be centrally managed from the HP Data Protector console.
In a Microsoft Hyper-V environment, backup administrators can install the online backup agent in two places: the child partition (virtual machine) and the parent partition (the root partition on which the virtualization stack runs). Both sit logically on top of the hypervisor. Backup administrators can choose from the following options:
• Traditional online backup agents inside the virtual machine
• Hyper-V server-based snapshots
• Offline backups
• HP Data Protector Zero Downtime Backup (array-based snapshots) and Instant Recovery
Traditional online backup agents inside the virtual machine
Administrators install an HP Data Protector online backup agent inside each virtual machine (or child partition) and, while this approach delivers consistent backups, it impacts server performance.
In a Hyper-V environment, restoring data from a virtual machine is very similar to restoring it from a physical machine. The backup administrator browses for the desired objects in the HP Data Protector console and retrieves information with a few simple mouse clicks.
Microsoft Hyper-V server-based snapshots
Data on child partitions can also be protected by executing sever-based snapshots. In this case, an HP Data Protector agent is installed directly in the parent partition.
Hyper-V provides tight integration with the Volume Shadow Copy Service (VSS) framework, which enables snapshots to be taken directly from the parent of one or more child partitions. When the snapshot is complete, HP Data Protector backs up the snapshot image to the desired target.
As mentioned above, these backups will be application-data consistent. However, this method can still impact performance because all backup operations are being performed on the server.
When restoring server-based snapshots in a Hyper-V environment, the administrator uses HP Data Protector to restore one or more child partitions within a Hyper-V server. If the physical host goes down, you would also need to restore the parent partition—which contains Hyper-V configuration information. Partitions can also be restored to another physical Hyper-V server.
Hyper-V only allows restoration of an entire child or parent partition, regardless of whether you’re restoring to the same or a different physical server. During a restore, Hyper-V stops the partition being restored (if it’s still running), deletes it, and restores the entire partition from the desired backup target. After the restoration, the partition is always offline and needs to be restarted by the backup administrator.
To use the server-based snapshot method described above, the child partition needs to be online and based in a VSS-aware operating system (OS). If the child partition is offline, or based on a non-VSS-aware OS, the Volume Shadow Copy Requestor component is unavailable. In this case, Hyper-V only allows offline backup. HP Data Protector automatically detects the configuration and, as appropriate, executes a saved-state— or offline—backup.
When HP Data Protector initiates a backup, one of two things happens. Either Hyper-V identifies the child partition in the state in which it was saved prior to being shut down or powered off by the backup administrator, or the VSS writer suspends the child partition before it begins the backup.
The backup process then proceeds as it would in traditional suspend mode in a physical environment, with the data being backed up to disk or tape prior to restarting the application.
In this case, backed-up data is only crash-consistent with the application once restored, because the application was “frozen” at the time of backup.
To restore data, you simply use the HP Data Protector console and follow the same restore process described for Hyper-V server-based snapshots.
Array-based snapshots (HP Data Protector Zero Downtime Backup and Instant Recovery)
All of the above methods consume Hyper-V host resources to run the backup, which impacts server and application performance. In addition, the offline backup method yields only crash-consistent data. HP Data Protector Zero Downtime Backup (ZDB) is currently the only way to avoid server performance degradation in a Hyper-V environment and recover to the second of your choosing—all from one interface.
In a Hyper-V environment, you have two options for installing HP Data Protector ZDB. You may install the ZDB agent in the parent partition, which would oversee execution of snapshots on all the child partitions in that server. Or, you may choose to install the agent in one or more child partitions and thus utilize ZDB functionality for specific applications only. As with the server-based snapshot example above, in this case all the child partitions must be running in Microsoft Windows and the applications must be VSS-aware. In addition, HP Data Protector maintains snapshots on the disk array for use by HP Data Protector Instant Recovery (IR).
For business- or mission-critical applications, your backup software must be able to automate the creation and recovery of snapshots on virtual servers. And, for some applications, the ability to recover to the minute – or even the second – between snapshots is crucial. HP Data Protector software can access application transaction logs (where all of the application’s transactions are recorded before they’ve been processed and/or written to data file) and fill in the data gaps between snapshots. HP Data Protector then synchronizes the application with the data, and restarts the application to the exact point in time you need. You simply enter the hour, minute and second into the Data Protector console.
In mission-critical environments, the ability to automate this process without utilizing a scripted or manual solution is critical. Both can leave room for errors and sap valuable time.
For more information on virtual server protection, check out our blog: Protecting VMware environments with HP Data Protector on the HP Software & Solutions Data Protection Community. Or watch the HP Data Protector software webcast: Best Practices for Data Protection in a Virtual World.
Don’t miss a thing. Get the latest information on HP Data Protector software.