Feature request - backup/retention per virtual disk

[DavorS]

Hi, i'd like to request feature that might prove usefull for file servers that host different data.

We often see file servers with multiple virtual hard disks, each specified for different data. For example, some might contain sensitive work-in-progress data that needs to be saved every two hours within the working hours with and need to have 7 days of restore point, others contain mostly static data that has to be backed up monthly, other disks have to be backed up mid-day... etc

Now we have to specify different jobs and exclude all disks but the one we need and set the retention and schedule parameters, and also we can't make use of built-in deduplication as it is done on per-job basis.

Having an option to define those parameters inside the same job on per-disk basis would be awesome. It doesn't have to be within a backup job, it can also be a specific job type like "File server backup job".

[veremin]

Hi, Davor,

But how you're going to address inconsistency between disks in case they are backed up at the different time? (restore scenarios)

Thanks.

[DavorS]

Well, one approach could be to mount the latest restore point for all disks (since in most cases you return from your latest backup) - if you're restoring from a complete virtual machine failure. Other approach might be an extra step in selecting a disk from which you want to restore data, then selecting a specific recovery point (like an extra step in the classic guest os file restore wizard).

That's why i suggested it could be an additional backup job type (like backup copy, etc...)

[dellock6]

Davor, and what about potential inconsistencies that can arise when having different disks at different restore points? I think for example a database with db files in one disk and logs in another...
As per one the reason to skip the backup of less used disk, this is automatically done by CBT, as a disk in not changing its content during time, even the incremental backup of the entire VM will not copy unchanged data, while at the same time guaranteeing the consistency of the VM itself.

[DavorS]

Luca, that's why this is a feature intended for a file server only - there is no relation between data except that it's presented through same file server. Ofc it wouln't be smart to do it on data that has to be consistent CBT is nice, but won't help you if you have a company policy that states different retention time on different types of data (disks) and RPO's.
I do agree that cbt will help to some extent on disks that don't change as much, but i've seen people dump data on disks that needs to be backed up once a month for archival purposes.

[larry]

Most servers I don't need the system drive after a few days its the data drives I need to retain. now I retain both system and data because of the overhead of keeping many jobs and the risk I miss a drive if I split it. I do have one job that backs up a SQL database but no other drives that I placed in its own job because I need to retain every day for 7 years.