Best Approach for transformation from ext4 to BTRFS including replacement of rsync-backups

[milkytwix]

@dalto

Err…first @ is just a convention. You don’t actually need to name your subvolumes with @ unless you want to.

Second, snapshots don’t descend into nested subvolumes so this shouldn’t be an issue either way,

Wait what… That makes things WAY more easily. So I just create a subvolume via “sudo btrfs subvolume create /home/milkytwix/Games/” and I am good to go, snapshots of /home will not secend into it and everythings fine? Really that simple? Asking for a friend, ofc…

@dalto

I don’t really understand what you mean at all with this statement.

Because I wans’t aware of the question (and answer above) I planned to mount “games” from the “outside”, so to speak. But this is not required.

@dalto

If you want to use Btrfs Assistant you definitely shouldn’t create manual subvolumes for holding snapshots. That will definitely make it harder.

Understood, again: Maiking things easier!

@dalto

No, it works fine.

Well… I will go with the first, because… no descending into nested subvolumes

@dalto

They will always be separate. You are taking snapshots of a subvolume. How you nest or mount them won’t change that.

Well again: no descending makes things WAY more easier

@dalto

Above you mentioned you were going to use Snapper. Snapper and Timeshift are alternatives. You should not use them both. Timeshift can only take snapshots of @ and @home. Nothing else.

Snapper can take snapshots of any subvolume.

Well, I definately want this:
https://discovery.endeavouros.com/encrypted-installation/btrfs-with-timeshift-snapshots-on-the-grub-menu/2022/02/
And because snapper is included in BTRFS-Assistant I intended to use it as well… can I have grub snapshots without timeshift? Including auto-generated ones after updates? Because I just love that feature.

@dalto

@ is convention that generally refers to the filesystem mounted at /. If you have a second disk, I would not recommend creating a subvolume named @ on it. That would be confusing in a lot of ways.

Don’t get hung up on the whole @ thing. It is just a name. For example, none of my installs have subvolumes named that way.

Understood! The next questions are just out of curiosity:
Why don’t you use them? How do you avoid them being created during install?

@dalto

Yes.

Thx!

@dalto

Is your TrueNAS box using zfs? If so, you could use btrfs for you main drive and zfs for your data drive. zfs supports snapshots and snapshot replication as well. It is a little harder to configure though.

A simpler option would be to use snapshots locally and deduplicated borg backups on your NAS. Those backups are nearly as space efficient as btrfs snapshots. The disadvantage is they are a lot slower to create.

I already use snapshot replication for the backup of my NAS, that won’t be an issue. My original plan was to use ZFS on root (Remember? Another "Help, migrating from Manjaro" topic - #2 by dalto ), so I will stick to ext4 for the time being on my non nvme drives and slowly convert them to ZFS.

@dalto

Personally? I use a combination of snapshots, snapshot replication, borg and rclone for a remote replica of my encrypted borg backups.

Any important data is stored at a minimum of 3 times and at least 2 geographically disparate locations. In many cases I have 4 copies of the data and also employ raid at most of the storage locations.

The first part was the interesting (I was asking for the technical implementation, not the backup concept, sorry for asking in an unprecise manner). My backup concept is: Data on NAS, local Data is copied to NAS, together with backups of /home and stuff. NAS gets nightly backupped to secondary NAS (which resides in the garage), and Monthly backupped to a Harddisk which rotates to my sister.
Coming back to “how you do it”: I will read into rclone and borg and try to find something suitable, thank you VERY much!