Veeam & Cloudarray (Twinstrata) cloud backup

[toon.dillen]

Hello

I'm testing a way to bring my Veeam backups outside the office into the cloud. One solution that I'm testing is CloudArray from Twinstrata.

What basically happens with this solution, is that the CloudArray software provides an iSCSI disk to the Veeam backup server. This iSCSI disk has a local cache and then sync's to the cloud in the background (different providers possible).

When I run a full backup, I have around 3TB. I use forward incremental backups and every day I have around 30GB of changed data. We have a 10Mbps line upload, so uploading 30GB to the cloud takes less then an hour every day (so no problem).
But when you use forward incremental, it is necessary to periodically make a synthetic full backup or active full backup. When I use synthetic full backup, I understand that - when it is time to take the synthetic full backup - Veeam takes an incremental backup from my VMWare environment and then constructs a new full backup based on the previous full backup and the incrementals in between. This is good because the VMWare environment is not used for this action.

But for CloudArray, there is a completely new full backup on his cache disk and it starts uploading this full disk again. I have 3TB and can only upload 600GB/week, so that is not going to work.

I can use the forward incremental in combination with the Active Full backup and then for example choose to only have a full backup twice a year and forward incrementals in between.

Would this work and can I be sure that I have a good backup chain? Does Veeam check this chain every time a forward incremental is taken? What is the recommendation on the maximum number of incremental backups in between 2 active full backups?

Any recommendations on this? Or other solutions that can accomplish my goal of backup up to the cloud?

Thanks
Toon

[chadtandy]

I'd do reverse incremental, this way you'd always have a current full with rollbacks.

-Chad

[toon.dillen]

Hello Chad

Thanks for your feedback!

Ok, but if you do reverse incrementals, Veeam constructs always a new full backup. This is not going to help me because then I will have even more changed data on my CloudArray disk -> so it would be even more impossible to get it into the cloud.

Or am I making a mistake somewhere?

Thanks
Toon

[Gostev]

Hi Toon, actually that is not correct. In case of reversed incremental, Veeam in-place updates the existing VBK file, instead of creating new one. Thanks!

[dellock6]

Uhm, maybe the problem is about the VBK file changing name every day. I do not know how twinstrata handles this file when it changes its name, anyway there is a workaround:
- open regedit on Veeam Backup Server
- browse to HKEY_LOCAL_MACHINE\SOFTWARE\Veeam\Veeam Backup and Replication
- Create DWORD = DisableVBKRename
- Set Value = 1
- Restart Veeam Backup Server
- Perform a new full backup and then the filename of the VBK will not change anymore

In this way probably twinstrata can better handle the full backup even with incremental. For sure this works on rsync.