Hi, thanks for sharing!

I just wondering in your second table: Could it be that for these two lines Benefit of Fast Clone (and storage IO) should be exchanged? For my understanding, you get Fast Clone benefits from synthetic operations that forward incremental uses and forever forward incremental does not.

@eprieto: Just two additions from my side:

• Isn’t the speed of backup for all three types the same? Only the runtime of the job will differ?
• regarding the backup corruption; if we use ReFS or XFS I would say a file/block corruption could result in the loss of the complete backup chain, not only the latest state. On those filesystems only an active full could survive (in theory)

Apart from that I really like your comparison and will bookmark it for the future :)

@vNote42: I’m not sure if I missunderstand your reply, but forward incremental doesn’t use synthetic operations at all?

@regnor : Yes forward incremental does not use synthetic operations. Therefore it does not benefit from Block Cloning. But when you use periodic synthetic fulls during your forward incremental chain, you benefit from Block Cloning. Instead if you use periodic active fulls, you do not benefit from Block Cloning.

For my understanding, “forever forward incremental” does just increments; “forward incremental” uses some kind of full: synthetic or active. Because of this understanding there seems to be something wrong in the second table. Or, of course, my understanding could be wrong.

@vNote42: You’re absolutely right and now I see what’s confusing here. I think the table (or @eprieto) meant an incremental run of a “forward incremental” job, which doesn’t benefit from fast clone by itself. If a forward incremental is scheduled to do a synthetic full or active full the result differs. Perhaps it would be easier if the table said “Forward incremental - Synthetic Full” (yes) and “Forward incremental - Active Full”(no).

At the end of a forever forward job, there is a merge operation where contents of the oldest VIB file in the backup chain is injected into the VBK file. This can be offloaded with Fast Clone.

This is a great question @regnor. There are some differences. When using reverse incremental, there is some copy on write operations, but this happens during the actual backups. Forever forward incremental has a merge operation, but the job only kicks this off after all backups are done.

Yes, you are absolutely right. But even reverse incremental benefits here from Block Cloning.

100% agree, thus why “reverse incremental” is marked with a green check box on the second screenshot in the original post.

In case anyone prefers some official source, the user guide is pretty clear:

Veeam Backup & Replication uses Fast Clone for the following operations:

In backup jobs: merge of backup files, creation of synthetic full backups, transformation of reverse incremental backups and compact of full backup files.
In backup copy jobs: merge of backup files, creation of GFS backups (synthetic method) and compact of full backup files.

Hi @eprieto I'm new one here in community thanks for sharing this information.