zfs-vm is an utility for syncing ZFS snapshots of VM/container filesystems.

Commands

• List zfs-vm snapshots on specified host (localhost by default).

    list [[user@]host]
  

• Pull snapshots from specified host and put them to "local-parent-fs".

    pull [-n name] [-d local-dest-fs] [user@]host
    -n - pull only snapshots with the specified streamline name
    -d - specify destination parent fs on remote
  

• Push snapshots to specified host and put them to "remote-parent-fs".

    push [-n name] [-d remote-dest-fs] [user@]host
    -n - pull only snapshots with the specified streamline name
    -d - specify destination parent fs on remote
  

• Rebase collection of datasets by creating consolidated dataset and creating clone for each source dataset based on this consolidated dataset.

    rebase -n name [-r] [-d] [-f] [-s suffix] [dataset...]
    -n - consolidated dataset snapshot name (xyz@snap) which be base for re-based datasets
    -f - destroy target consolidated dataset if already present (non recursive)
    -d - remove original data sets (non recursive)
    -r - replace original datasets with re-based clones
    -s - suffix to add to original dataset name to cloned datasets
  

Options

zfs-vm.py has the following global options:

-d	debug
-n	no-op
-s	use sudo on the remote side
-v	verbose

Examples:

• List snapshots on localhost.

    zfs-vm.py list
  

• List snapshots on remote-host.

    zfs-vm.py -s list fjoe@remote-host
  

• Push all snapshots to remote-host.

    zfs-vm.py -s push fjoe@remote-host
  

• Push all snapshots to remote-host and put them to tank/vm.

    zfs-vm.py -s push -d tank/vm fjoe@remote-host
  

• Pull all snapshots from remote-host.

    zfs-vm.py -s pull fjoe@remote-host
  

• Rebase dataset (file systems) collection by creating consolidated data set and replcing original datasets by it's clones instaltiated from consolidated data set and keeping backup of original ones.

    zfs-vm.py -s rebase -n tank/vm/merged@today -r tank/vm/vm1 tank/vm/vm2 tank/vm/vm10
  

Push/pull

If the snapshot is completely missing on the receiver side the full stream is sent, otherwise push/pull identify the minimal incremental stream sequence required to sync snapshots.

TODO

1. zfs-vm.py should support recursive behaviour for pushing/pulling more than one FS.