ScaleOut Repository or Windows Spanned Disk?

[timmi2704]

Hi experts,
usually, our backup storage system presents several 64TB volumes to our backup server.
In Windows, these volumes are spanned as a single disk. In Veeam, we use this disk as a backup repository.
Now with ScaleOut repositories available, I ask myself, whether it makes sense to keep this volumes separated and rather use ScaleOut repository functionality to aggregate the different disks.
I think this would maybe give a bit more flexibility. I do not expect any performance improvements.

Any thoughts on this? Any disadvantages?
Thanks

[PTide]

Hi,

Any thoughts on this? Any disadvantages?

Well, all depends on what types of jobs you have in your environment. Besides management convenience scale out backup repos have some limitations that you should be aware of. For example, you cannot use SOBR as a target for configuration backup job, replication jobs, endpoint backup jobs. Please see the article I referred to for more details and let us know if any questions persist.

Thank you.

[dellock6]

In addition, you should evaluate how are your jobs configured today, and if that's the reason why you are now grouping volumes using spanned volume. For example, if it's because a single full backup is bigger than the single volume, this is a problem first of all because 64 TB is bigger than what we usually suggest as a manageable backup file, and second because the single spanned volume is still a single failure domain. If just one volume breaks and the file is spanned over multiple volume, the entire file is lost.

I'd first of all switch to per-vm chains to keep backup files under reasonable size, and leverage multiple streams. This is an enhancement that can be reached without SOBR. Then, I'd also personally use SOBR to keep the different volumes separated, and thus be able to improve the management of those volumes: they can be put into maintenance, removed, without impacting the entire spanned volume.

[timmi2704]

Thanks, Pavel and Luca, your answers are greatly appreciated!
I think I will change to SOBR as soon as the update is installed (currently waiting due to some problems).
I like the flexibility, although the storage is still the single point of failure as all extends will be located there.
I definitely need to watch out for the limitations, Endpoint backup is in use.
Biggest backup file size is 7TB, but I will change to per VM chain anyway.

[dellock6]

If the volumes all come out from the same physical array, you should also look at which controller is managing a volume, otherwise you are running different volumes (so different SOBR extents) using the same physical resources. It can be done, but for example transforms will not see the advantage that SOBR promises.

[timmi2704]

Yep, that's true.
The volumes are distributed over the different controllers, but these aren't the bottleneck here
I'm rather suffering from slow disks, but that's another story.