Remote Copy question

[Bobby Gillette]

This is pretty much a no-brainer, but being new to LeftHand it begs asking. If I were to reduce the number of Remote Copies of a volume could I conceivably increase the amount of usable space on my cluster (ie we are currently keeping 3-5 remote copies on each cluster, if reducing that to 2 across the board for each volume)?

Any thoughts?

[Gauche]

Keeping less snapshots will take up less space. But how much will depend on the degree of change in between each snapshot. That is not usually a large number.

[Bobby Gillette]

I realize that on local snapshots, what I'm specifically referring to are remote copies, which I understand to be at minimum a full copy of the volume, with additional remote copies acting as incremental snapshots.

At minimum, if I were to reduce the number of remote copies (or remove them altogether since they aren't off-site and we're strapped for space), wouldn't I at least reclaim the amount of free space equal to full copies (the remote copies) of each affected volume?

For example, I have one volume that's ~600GB, meaning it's initial remote copy is also 600Gb. Deleting the remote copy and associated schedule should at least net me 600GB of free space on my SAN, correct?

If I'm off-base please enlighten me :)

Thanks in advance!

[Gauche · Accepted solution]

If you just delete some of the versions of a remote copy you'll free up a little space depending on how many changes were stored in the remote copy. If you delete the entire remote copy you will free up significant capacity, but probably not equal to the size of the volume. In your example, if you have a 600GB volume, that you write 100GB to, then remote copy, you only have a remote copy of 100GB and deleting the remote copy will free up 100GB (not 600GB). RC does not copy empty space in the capacity of a volume, so the size of the volume does not matter as much as how much has been written to it.
I hope that helps some.

[Bobby Gillette]

Exactly the answer I needed. Now if only there were an accurate way to know how much space could be reclaimed by reducing either the number of snapshots and/or remote copies... since we're doing thin provisioning it's somewhat fuzzy math, but given our current situation, it may be warranted to free up as much as possible.

I'll open another thread on a similar question you may be able to comment on as well. Thanks!