Hi, when you want to use zfs send/receive to make incremental backups, do you need to keep all the snapshots you're making the backups from around indefinitely?
I haven't found any documentation about how to deal with all the snapshots which would be created over time. Can they be destroyed once the backup is finished? A full backup took about 48 hours, so something faster is needed, and I don't want to end up with hundreds or thousands of snapshots by making new ones every day without being able to ever destroy them. The manpage is entirely confusing: ,---- | -i snapshot|bookmark | | Generate an incremental send stream. The | incremental Incremental in which way? | source must be an earlier snapshot in the destination's | history. It will commonly be an earlier snapshot in the I don't want to back up the destination, and I don't care about its history. It's not like I'd be modifying the backup in between the increments. | destination's filesystem, in which case it can be speci‐ | fied as the last component of the name (the # or @ charac‐ | ter and following). Huh? | If the incremental target is a clone, the incremental | source can be the origin snapshot, or an earlier snapshot | in the origin's filesystem, or the origin's origin, etc. `---- There is only one source, which is the current data I want to backup. Should I make an incremental clone on the destination machine? Basically, documentation says that such incremental backups are awesome because you get a 1:1 copy and only need to transfer what has changed after a previous backup as if you would use rsync, but that it's better than that and you can do it in like no time. It doesn't really say how to actually do that and what to do with all the snapshots, though. I also can only guess that enabling compression on the target FS won't work unless compression is enabled at the source, though it would be rather useful to have the backups compressed while the source is not. You could do that with rsync, though, but I don't know how to access the snapshot for that. So how does this work?
