HP P4000 - Replication to Off-Site Best Practice

[AnthonyWa]

Looking for some best practices, we have two management groups of P4000s, one is for Production and another is for off-site replication.  The production P4000s handle our VMWare Virtualized Enviornment. Each night we replicate the data from the Production P4000s to the off-site replication copy of the P4000s using the P4000 remote copy feature.

The biggest downside is that the Production P4000s have to maintain a copy of all the data in a snapshot.  So, even though we have 10TB of capacity on our production P4000s, we can only use 5TB, as we have to keep the other 5TB for snapshot.

Is there anyway around this?  Maybe through deduplication or some other technology?  Or another way to replicate the data without having to basically have double the storage on the production side?

Any ideas welcome, thanks.

[oikjn · Accepted solution]

sounds like you have the LUNs set as THICK provisioning.  Set them to THIN and then the snapshots will only be as large as the actual changed data.  There is generally no need to use Thick LUNs.

If you are really paranoid about over-provisioning, just create one LUN that nobody connects that is thick provisioned with some size buffer you think is appropriate and then if you ever do run into a full capacity problem, you can quickly delete it and get RW access to the luns back online. 

[AnthonyWa]

Ok, thanks, that got me going down the right path.  It's actually due to the Network Raid 10 that the volumes take up 2x the space, not the snapshots themselves.  thanks.

[oikjn]

yes, NR10 will use 2x the raw capacity.  A 5TB LUN should take up 10TB of raw SAN space when using NR10, but a snapshot of that 5TB LUN should only take up 2x the actual LUN delta over the time it exists starting at a minimum of 1GB (2GB for NR10). 

Are your volumes Thin provisioned or thick?  HP's best practice is THIN.