Question regarding Active Full Method for Archive Backups

[btmaus]

Hi everyone,

I've recently upgraded to Veeam v9 and I would like to try and make use of some of the new features to improve backup efficiency.

Currently when Veeam does the Backup Copy job transforms, the storage target takes a big hit in performance.

I have a Dell DR6000 dedupe appliance which I use for my backup jobs, lets call it "LUN A" on the appliance. Now on the same appliance, I have "LUN B" which is used by the Backup Copy jobs as their target.

I read about the new option called "Read the entire restore point from source instead of synthesizing it from increments". Would you advise that I make use of this feature considering my current environment?

[Gostev]

Hi, Glenn - absolutely, as this feature was basically designed to be used with deduplicating storage which we do not have custom integration with. Thanks!

[btmaus]

Thanks Anton.

Just to confirm, it doesn't matter that I have my backup job target and backup copy job target on the same storage appliance for this new setting?

I see the notes says the following:

To optimize the backup copy job performance on deduplicating storage appliances, you can enable the Read the entire restore point from source instead of synthesizing it from increments option in the backup copy job settings. Veeam Backup & Replication will copy data for the archive full backup from restore points on the source backup repository, transport it to the target backup repository over the network and write it to the archive full backup file. The load on the network will be higher but the performance of the deduplicating storage appliance will increase.

How is that different to a transform that was being done by version 8, as the read and write is still on the same storage appliance with this new setting?

[foggy]

Generally speaking, active full in this case is still just 1 read and 1 write (sequential), while transform requires 1 read and 2 writes for each block and these I/O are pretty random. Even though the amount of blocks to be processed in case of incremental backup is much less, active full might still be more preferable for your storage. You can run some test backups to verify.