FastClone support downloading from Capacity Tier ?

[DBMandrake]

Hi All,

After pouring over documentation and doing other research I can't seem to find a definitive answer to what happens with FastClone when downloading backups from the Capacity Tier of a SOBR to the Performance Tier, in the event of having to rebuild an appliance. If anything the documentation seems to contradict itself.

The scenario:

A SOBR where the performance tier is the built in (hardened) repository in the Veeam Software Appliance (which supports FastClone using XFS) and the Capacity Tier is Data Cloud vault.

In the performance tier space savings are achieved by Synthetic fulls using FastClone, so effectively only the difference between the fulls is consumed on disk.

Now consider the scenario that the appliance is lost and a new appliance is set up, a configuration backup is restored, which reconfigures the SOBR, connects to Data Cloud vault etc.

All backups exist in the capacity tier but none in the performance tier. Before resuming backups, if you want to avoid a new full you would need to download at least the active chain from the capacity tier first ?

Documentation has the following advice:

https://helpcenter.veeam.com/docs/vbr/u ... tml?ver=13

Veeam Backup & Replication uses the Fast Clone technology only if you download all offloaded backups back to the performance extents using the Download option. Fast Clone is not supported for the Copy to Performance Tier option.

This makes it pretty clear you should not use the Copy to Performance Tier option if you expect it to take advantage of FastClone, and that you should use the Download option.

However the download option lets you choose the active chain only, all chains including inactive, and if you choose to download only the active chain now, you can choose to download the inactive chains later as a separate step...

Is downloading all together equivalent from a FastClone point of view to downloading active chains now, and inactive chains later, or will the inactive chains not share any common FastClone disk space with the active chain in that separate download case ?

So you could either:

1) Download the active chain only, (faster) resume backups on top of that active chain (avoiding a fresh full, saving both disk space and backup time) and then accept that your older backups are cloud only (no 3-2-1)

2) Download the active chain only, resume backups on top of that active chain, and then download the remaining inactive chains after backups had resumed. (But would these share any disk space with active chain ?)

3) Download all chains (slow) resume backups on top of the active chain.

A single line in another document calls everything above into question however:

https://helpcenter.veeam.com/docs/vbr/u ... tml?ver=13

Full restore points downloaded from capacity tier to XFS performance extents do not use Fast Clone.

I am hoping that what this line actually means is that individual full restore points downloaded using the "Copy to Performance Tier" option would not use Fast Clone, not that any method of downloading including the download option would not use Fast Clone, as that would directly contradict the first quote from the other documentation page that says download does take advantage of Fast Clone.

So does anyone know the answer to the above for sure ?

Will "Download all" for a backup job to bring it back from the capacity tier to an empty performance take full advantage of Fast Clone as if the backup had always been there ?

Will downloading only the active chain using the download option and then later using the download option to download the inactive chains provide the same Fast Clone space savings ?

Thanks!