Renaming a Dataset, Impact on Replication and Snapshots

[Daniel Carrico]

I understand that datasets can easily be renamed using the shell and the following command:

zfs rename zpool/oldname zpool/newname

Simple enough. My question is what happens with pre-existing snapshots and how is the renamed dataset handled on the receiving end of a replication task?

Here's my setup. I have recursive snapshots enabled on Server 1 so that all of the datasets on my zpool are always included. I also have a replication task setup to mirror the entire zpool with snapshots to a remote server, Server 2.

If I rename a dataset on Server 1:

1. Will the snapshots for that dataset be associated with the renamed dataset on Server 1 following the rename? Or will they be lost?
2. Will my recursive snapshot task on Server 1 continue on with the renamed dataset?
3. Will the replication task also rename and replicate the renamed dataset on Server 2? Or does it see the renamed dataset on Server 1 as a new dataset, and simply copy it over like any other?

Are there any other potential snags I should be aware before I rename a dataset?

I'm running 9.10.2-U5 on both Server 1 and Server 2.

 

[melloa]

I'd create the new dataset using the GUI, so will be available there (new snapshots will need to be configured).

For ZFS send/receive, take a look at this old thread: https://forums.freenas.org/index.php?threads/zfs-incremental-send-on-recursive-snapshot.16712/

If you don't have the extra space, mv from the shell would be a solution to move the data. Not sure about the snaps.

As always ... back-up first!