Veeam & Deduplication Appliances

[stevil]

[merged into existing discussion]

I could have sworn I'd seen a thread similar to this, but I cannot seem to find it despite several searches.

Basically, we have a large IBM NAS that does its own dedupe. I've tested backing up to it using only changed block tracking, no compression, and leaving the NAS to do the dedupe. This appears to work OK, but I was wondering what other customers experiences were using a deduplicating NAS and how their jobs are set up for optimal space saving? I did toy with the idea of a full backup every night. In theory it would dedupe massively and I wouldn't need to rely on a glitch free incremental backup chain. I never did get chance to test this though.

[MB-NS]

Your description of the process is actually mostly correct, except your used space calculations. You are missing the fact that your target is deduping storage device. Thus, new VBKs will be taking little to no space. Remember, new VBK is created synthetically from data already available on device in previous VBK and VIBs.

Hello,

Exagrid deduplication is post-processed, not inline (like DataDomain).
So unless they implemented something different just for Veeam, I think one would need the space for 2 full, as they would be needed until the deduplication kicks in.

[Gostev]

Yes, my understanding is that ExaGrid has some sort of built-in "staging area" where the data goes before it is deduped and sent to "main" storage. But it is large enough, and this all happens transparently, so you don't really need to think in terms "I need to find disk space for 2 full backups".

[mcrespi]

Hello,

Marc Crespi from ExaGrid Systems here. Yes, Gostev is correct. When we size a customer we factor in the amount of useable disk required to store the most recent backup set in Veeam's native format in addition to the deduplicated copies of the required history. This does not result in a space requirement equal to 2 fulls but instead equal to 1 intact full and 1 deduplicated full for the most recent plus deduplicated copies of the the desired retention.

Unlike other deduplication appliances, having this intact version means that recovery of most recent backups will not require our deduplication engine to reconstruct the data. This allows us to provide very robust support for features like Instant VM Recovery, SureBackup, etc.

Sincerely,
Marc

[Gostev]

That's right, ExaGrid is basically ideal storage for SureBackup, because SureBackup jobs verify the latest backup. And even from perspective regular restores, in 98% of cases people need to restore data from the latest backup (and not some earlier point in time). So, definitely good to have it available on fast storage for instant VM, application item or file level recovery through vPower engine.

[dephcon5]

is there a Veeam white paper on how to setup veeam to hit any of these dedupe appliances as the target? I am currently pilot testing the Quantum DXi

[Gostev]

No special setup is needed from our side, just backup to a CIFS share on the device.

Dedupe vendors would typically recommend using Veeam in incremental backup mode, disabling compression, but keeping dedupe on. This way you will get best performance and dedupe ratio from storage perspective.

We have not completed formal testing with Quantum yet.

[jpaul]

I am doing a series of Blog posts on Veeam with the Exagrid and the mileage I'm seeing in a real customers environment.

Here is the first post, but there will be more to follow before the end of the month.

http://jpaul.me/?p=1585