Is it possible to create & rollback a snapshot directly on the pool level? Or does this need to happen on the dataset level. Basically:
I noticed behavior on my "new" FreeBSD laptop that I would currently describe as "unexpected".
My procedure:
When running the last (4th) command I expect the shell to report that the file ~/foo.txt does not exist anymore as I'd expect it to vanish when rolling back to the snapshot taken prior to the file's creation. However, the shell just happily prints
What (obvious?) thing am I missing here?
zfs snapshot -r zpool@mybackup
I noticed behavior on my "new" FreeBSD laptop that I would currently describe as "unexpected".
My procedure:
$> zfs snapshot -r zpool@mybackup
$> echo > "Hello World" ~/foo.txt
$> zfs rollback -R zpool@mybackup
$> cat ~/foo.txt
When running the last (4th) command I expect the shell to report that the file ~/foo.txt does not exist anymore as I'd expect it to vanish when rolling back to the snapshot taken prior to the file's creation. However, the shell just happily prints
Hello World
.What (obvious?) thing am I missing here?