The behavior you describe with SanSymphony is the same behavior you’d see with a NetApp FlexClone. In the NetApp world, defragmenting your LUNs is a bad thing.

Also, keep in mind that FlexClones are not the only way to do thin provisioning in a NetApp storage environment. We can create thin provisioned volumes straight out of the gate by specifying no space guarantees when creating the original FlexVol. I haven’t done this with VMware and VMFS partitions yet, but do plan to test this functionality soon in order to observe the specific behaviors in that environment.

Love the blog and all the technical goodies. I was wondering about the FlexClones. I have much more experience with Clariion and SanSymphony’s thin provisioning so this is more of a question due to my lack of NetApp knowledge.

When we thin provision disk using SanSymphony it works in much the same manner as FlexClones in the sense that only the blocks of storage that change are the blocks that are actually used. With this feature, we have the ability to over allocate disk resources much like your idea of using FlexClones. The problem that I’ve seen is that defrags or even Virus Scanning engines touch all the blocks inside of a volume even if they’re not consumed, rendering the thin provisioning worthless, as 100% of the blocks in the disk are touched during that time. I wanted to know if this behavior happens within the NetApp provisioned disk when using FlexClones. Does a Virus Scan or defrag cause the FlexClone to grow back to 100% of the original volume size?

Thanks again for all the great work you share with us.

Rob Daly