Backup Strategies for Large VMs with and without Deduplication

BackupChain offers many options that allow you to quickly implement several backup strategies you can use simultaneously.
If you are dealing with very large files, say a 1.6TB VHDX being the largest, you may need to plan for restore times.
An average modern hard drive and server will give you processing speeds of 50 MB/sec to 100 MB/sec without using RAID.
A 1.6 TB file would hence require 1677721.6 MB -> 4.7 to 9.5 hours in order to restore, regardless how it’s going to be restored.

Using an external hard drive without backup processing (no deduplication, no compression) has the advantage of being able to boot VMs directly from the external drive when necessary.
This strategy saves you from having to restore for 5-10 hours for a 1.6TB VM.
However, as each strategy has a downside, this strategy requires 1.6TB for each VM backup; hence, you can’t hold on to too many backups. If you wanted to be able to restore files from a week back you would need 7 * 1.6TB = 11.2 TB of storage for just one VM!

Naturally most users can’t afford that and opt for a hybrid system.

Here’s how you would set it up:
#1: Set up a task to write a full, uncompressed, unprocessed backup to a dedicated hard drive every night. This way you can always go back and boot the VM instantly when needed. However you need enough space for at least 2 full backups on your hard drive as BackupChain won’t delete the old backup until the new backup completes.
#2: Set up a second task to run afterwards using deduplication and medium compression, targeting a different hard drive. This task should be configured to keep a week or more of backups. You can afford this storage-wise since it’s deduplicated and compressed. This strategy uses roughly 50% of 1.6TB data for the initial full backup (assuming the VHD is in fact fully used), and then about 5% for each daily increment.

Basically you would use Task #2 for long term recovery storage and task #1 for immediate recoveries when a complete restore is needed to be performed quickly.
Either way, it looks like even for task #2 scenario a 2.7 TB hard drive won’t be large enough to hold all that data.

You can either look for larger USB drives or set up a small Windows network server with several 2.7 TB drives combined into a software RAID. That’s a great and economic way to add storage.
NAS boxes that aren’t using Windows and NTFS internally may give you issues with large file sizes, lengthy network connections, and deep paths beyond 240 characters.
Every specific NAS box is different so it needs to be investigated on a case-by-case basis if it’s suitable or not. Those NAS systems using Windows Storage Server OS are more expensive but you can be certain you have Microsoft standards compliant storage that won’t give you issues.

Tags: , , ,