On Jun 9, 2009, at 9:46 PM, mark roles <[email protected]> wrote:
I'm interested in the snapshot and cloning side of ZFS.
Whilst I'm happy with the way that this works at the share level i.e.
snapshot share
clone share to a new share
nfs mount (or somesuch) to the cloned share
I'm wondering whether it's possible to snap / clone at the file /
dierctory level so that I don't have to create additional NFS shares
to mount. If it's not possible at present is this something that
could potentially be investigated or is the design of zfs snapshots
such that this will not be possible (i.e. all snapshots under a
share are accessible through a single .zfs directory)
I know this is something that NetApp filers are able to do and it is
particularly usful in the VMware (VDI) type of environment.
VMware is limited in the number of mounts and creating a mount for
each VM (so I can have independant VM snapshots) makes manageability
difficult at best
You can do that.
For example I have a zpool md1000-1 with a zfs dataset vm_data1
mounted on /export/vm_data1 and shared via NFS to my ESX hosts. I
create another dataset called md1000-1/vm_data1/vdi_master and it will
be under the share and visible to the ESX hosts. I then take a
snapshot call it vdi_mas...@host1, then clone it to md1000-1/vm_data1/
host1 and it too will be under the share and visible to the ESX hosts.
I can then keep snapshotting and cloning vdi_master to make as many as
I need without adding any additional NFS shares.
You can then send/recv updates of the vdi_master to all it's clones
whenever there is a major upgrade.
-Ross
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
