On Friday, July 12, 2013 11:29 AM, Geoff Nordli wrote:
On 13-07-11 07:25 PM, Jim Klimov wrote:
On 2013-07-12 02:43, Geoff Nordli wrote:
The FSRVP could initiate a snapshot from the windows machine. Right
now
the windows machine consumes snapshots via the previous versions from
snapshots initiated on the zfs side.
I see. Well, it is different from what I meant, and a standardized
protocol is good for interop and interface (sometime...) ;)
On a side note, you can always make ZFS snapshots via filesystem
interface "mkdir $dataset/.zfs/snapshot/$snapname" over both NFS
and CIFS (at least kCIFS) as well as locally, as long as your user
has proper ZFS delegated permissions. Removing is trickier, since
tools insist on only deleting empty directories, and somehow they
fail to delete contents of a read-only snapshot... bummer ;)
That is very cool Jim, I didn't know I could create a snapshot that way.
Time to rewrite the scripts. This makes it way easier to manage
snapshots when you have over 2k snapshot per dataset. I've already
resorted to referencing one dataset's snapshots with ls since zfs list
takes forever. Thanks for letting us know about this. zfs just got even
better.
_______________________________________________
OpenIndiana-discuss mailing list
[email protected]
http://openindiana.org/mailman/listinfo/openindiana-discuss
