Hello,

I already know all the limits of having 1 zfs pool for each drive but having a cloud backup i care only for the snapshot (due to fat fingers and wrong deletes), compression and corruption warnings

The issue is i cant understand how to perform a dataset share-level

If I have a folder that gets replicated on more disks should I manually create a dataset for each disk? And what happens if i gather on one disk the files, each disk snapshot will not delete what we moved away, right?

I saw many scripts about auto-snaps but they grab as source only one pool and only one dataset in a pool/dataset fashion

Am I missing something?
Thank you all

1 hour ago, mich2k said:

If I have a folder that gets replicated on more disks should I manually create a dataset for each disk?

Yes.

 

1 hour ago, mich2k said:

And what happens if i gather on one disk the files, each disk snapshot will not delete what we moved away, right?

Not sure what you mean exactly, but snapshots will have all the files from the dataset/disk from the time it was taken.

 

You create you own script as needed, I use btrfs snapshots to backup all arrays from my servers, disk by disk, so it takes a snapshot from each disk and sends it to the backup server, I only have one share per disk, but you can have a script snapshoting multiple shares.