On Thu 28 Mar, 2002, Brian Tao taob@risc.org wrote:

On Thu, 28 Mar 2002, Steve Losen wrote:

...

Does ONTAP link the snapshotted blocks back into the live filesystem or does it just copy the file from the snapshot creating a new file in the live filesystem?

I believe the former is correct (if the operation is analogous to

volume SnapRestore). The advantage here is that I can now tell my DBAs they can roll back selected .dbf's, each 2GB in size, in a matter of seconds. How long would it otherwise take you to copy each 2GB file out of snapshots over NFS? 2 minutes? 3 minutes? :)

Key advantage isn't speed, it's viability.

On full filers a snaprestore in-place (by linking blocks from an old consistency point back into the live structures) means your very large file only takes space equivalent to its contents, which can make the difference between possible and not possible.

This is another feature I asked for years ago, back when we were moving from FAS1400s through F[23][123]0s to F540s. My how time passes.

-- Brian Tao (BT300, taob@risc.org) "Though this be madness, yet there is method in't"

-- End of excerpt from Brian Tao