clone / copy lun inquiry

Question

I'm new to VNX and am not the most experienced SAN person. Up until recently we had a different storage vendor, but we have migrated to EMC VNX. We have a server that we clone a lun on and then present it to another server for development. Once we clone it and present to the destination server there is no relationship between the source and destination lun anymore. We don't keep them in sync or resynchronize. The next time we need to refresh the data in development we go through the clone process all over again. I was wondering what the procedure in VNX is for doing this. From what I'm reading it seems that we want to do a fractured clone.

I have a few questions about the fractured clone:

1) Is this the appropriate feature to use on VNX for what we are trying to do?

2) Can a fractured clone reside in the same storage pool as the source lun? In our previous SAN environment our clones were in the same pool. We don't plan on having the cloned lun being synced with the source. We may want to resynchronize at some point because we want to refresh the data and I see that is an option with a fractured clone.

3) When you fracture a clone what's the relationship between the source and destination lun? Is there still a clone private lun when you do a clone fracture because I don't think this would be necessary to have if we are just looking to get a copy of production into development?

4) I'm looking at the VNX Unisphere Help > Fracture a clone - On Step 2 it says to stop I/O on the source LUN. We'd like to be able to clone a lun without having to take production down. That is how we used to do it in our previous SAN, why does it say to stop I/O because we'll have to get an outage window every time we want to refresh our development environment?

Thank you for your help.

dynamox · Accepted Answer

1) sure, you don't need to 'terminate' clone relationship and perform full sync every day. All you need to do is is fracture the clone and present it to your development system. When it's time to 're-sync' the clone, you un-mount the clone from target host, synchronize the clone again (it will be incremental), fracture it and present to development system again. And what i mean by presenting is you don't have to actually remove the clone from storage group. Different operating system handle it differently, for example on a Linux box i like to unmount file system, export volume group. Once when the clone is fractured i re-import the volume group and re-mount the file system. 2) Does not matter where the clone resides, obviously reading and writing to the same physical spindles that make up that pool will not be the most performance efficient, but it won't stop it from working. I know you mentioned that you are using clone for prod to dev refreshes but if you ever decide to use clones for backup purposes (say before it gets sent to tape) then i would recommend you place clones in a different pool. If something ever happens to the production pool, you still have your clone pool available for quick BC. 3) clone private lun is tracking changes when source <-> clone are in fractured state. CPL does not contain any user data (like RLP with snapshots) it is simply tracking which blocks are changing. My CPL are 2 x 10G LUNs and i am daily cloning a 15TB oracle data warehouse database. 4) It depends what you are trying to clone and how that application can recover. If you fracture a clone 'on the fly' you don't have 'application consistent' copy of the data. Your application could have a tons of data into server's memory that was not committed to disk so when you present that data to a dev system ..that system needs to be able to detect that 'hey, something is missing but that's ok ..i can recover'. For example when we clone our Oracle DB, we first put it hot backup mode, once it's in backup mode i fracture the clone and only then take DB out of hot backup mode. Again you don't have to take your application offline as long as it can recover from 'inconsistent' copy.

rgcda · Answer

Thank you for you answers.

So if you are just creating a clone of the source, fracturing it, and presenting it to the clone to the destination is there a need for the CPL? I guess I'm not totally clear on that yet.

dynamox · Answer

in order to use clone functionality on VNX you must configure CPL, regardless if you fracture and immediately delete the clone relationship or fracture and leave the relationship in place until the next refresh cycle. They don't have to be big, mine are 10g each and probably overkill.

1) make sure Fast Cache is not enabled on CPL LUNs.

2) Need two CPL, one for SPA and one for SPB

3) CPLs have to be traditional (flare LUNs), you can't use pool LUNs.

dynamox · Answer

What are the basic features of Clone Private LUNs (CPLs)?

chandlerbing1 · Answer

I thought CPL are only used for traditional LUNs?  If you are using VNX Pools do you still need to create the CPLs?

dynamox · Answer

yes

VNX

clone / copy lun inquiry

Was this post helpful?