De-duplication of blocks, does it work between full and Synthetics?

[Mani]

We've a full backup every 2nd Sunday which we retain for 2 weeks, followed by increments across the week and a Synthetic full on Saturdays. We end up with lot of storage used even with Veeam deduplication enabled. I want to understand if there is no other deduplication (Windows Server or Appliance); does the blocks between the two VBK (Full and Synthetic) deduplicate before written to the backup repository?

Please advise, thanks!

[Dima P.]

Hello ans welcome to the community Mani.

I want to understand if there is no other deduplication (Windows Server or Appliance)

Any chance you are using any deduplication storage as repository for your backups?

does the blocks between the two VBK (Full and Synthetic) deduplicate before written to the backup repository?

Deduplication works within the backup and not across multiple backups: during the backup job run two identical data blocks wont be placed into the same backup file. Check this article as to know more about deduplication options. Another good read is Deduplication and Compression article from Veeam Best Practices Guide. Cheers!

[Mani]

Hi Dima,

Thanks for your response.

Our customer isnt using any deduplication storage appliance, JBODs only.

So deduplication works only across individual .vbk files and .vib files?

What happens when synthetic full is created, does Veeam deduplication look at the gold copy (.vbk where .vib gets inserted ) and rest of backup chain, not inserting any duplicate blocks in the synthetic full?

I read about per-vm-backup chains on your 2nd article (thanks for sharing), does that increase dedup efficiency if enabled vs disabled?

Thanks

[foggy]

So deduplication works only across individual .vbk files and .vib files?

Correct.

What happens when synthetic full is created, does Veeam deduplication look at the gold copy (.vbk where .vib gets inserted ) and rest of backup chain, not inserting any duplicate blocks in the synthetic full?

Yes.

I read about per-vm-backup chains on your 2nd article (thanks for sharing), does that increase dedup efficiency if enabled vs disabled?

Since each VM is stored in a separate backup chain with per-VM setting enabled, dedupe efficiency is decreased in this case.

[Mani]

Thanks so much for the quick responses!

We can close this thread.

Cheers!