Restore of a deduplicated volume cannot be started

[LeventePeres]

Hey Guys,

I have the following problem. After a complete RAID meltdown, I need to do a complete tape-to-file restore to a volume that uses windows deduplication.

The available free space is limited to 64TB - but the complete restore data size is well over 110TB. When the restore is done, the files will occupy about 55-57 TB.

Since we don't have an automatic changer, the restore will be a multiday process, during which time deduplication can keep reducing the space between tape changes. This has been considered, and in fact, we're counting on it.

Right now we've already done some restores, and the free space right now is around 37 TB.

Unfortunately, Veeam won't let us start the next restore, because when it gets to the point of choosing which server to restore to, and where, it will simply report that there's not enough space available, and there seems to be no way of overriding this.

This is not a VMDK backup - it's old fashioned file-to-tape backup. Also, it's an incremental backup so it needs lots (dozens) of tapes, because we keep the block backups on the same sets to avoid having to change them several times every day. (I know...)

In order to keep the restore time managable, we need to restore larger (20-30TB at least) chunks, otherwise we'll still be doing this by Christmas...

There's no way to get temporary extra space. We don't have any such storage that could handle 100+ TB undeduplicated.

Is there any way to force veeam to ignore the free space and do the restore regardless? A reg setting perhaps?

Or maybe is there a way to temporarily report more space to veeam agent on the target server?

Thank you guys for any ideas.

[tkolstad]

Hmmm....that's a hard one. My fix would be to get some temporary storage to add, but as you said, that's not an option.

[david.domask]

Hi Levente,

No registry values or anything here, and I'm not sure it would be a good idea -- this is Windows Deduplication correct? Since it's a background job here, the data will be full size until the optimization job runs so even if we were to trick the software, I'm not confident this would help as probably there wouldn't even be enough space for the chunkstore to deduplicate the data.

I do wonder if Entire Tape restore would help you here? I understand you have multiple tapes for a given backup set here, but the Entire Tape restore is just a "tape contents dump"; even with LTO9 tapes, you should be able to get the contents onto the deduplication volume and then run the optimization job -- would the dump + deduplicate process perhaps work here? Because it will just dump the tape contents, it means that at most you're needing space for the raw capacity of the tape + potentially a little extra if compression when writing to tape was used.

If not, then the only thing I could advise is try to find any way to expand the storage, but I understand that finding an extra 40 TiB is not necessarily an easy task.

[RobTurk]

There's no way to get temporary extra space. We don't have any such storage that could handle 100+ TB undeduplicated.

Hi Levente,

Not sure where you are located, but have you considered renting a storage system, or even getting temporary equipment in from eBay?
A refurbished LFF storage server like Dell R740XD2 or an HPE Apollo goes for less than $2000 on eBay. Fill with 8TB disks and for ~$4K you have all the temporary space you need.
Perhaps that weighs up against the cost of you and your colleagues spending time juggling data within the current constraints.