I am recovering a Storage Spaces array that is 12.6TB in size. Using ReclaiMe's Storage Spaces Recovery to do it.
It's a long story, but essentially I was running in a MS unsupported way, presenting each drive as its own RAID 0 to Windows, then had a RAID controller fail. 29 drives, that when I hooked back up, because they were set to Windows through a RAID card, Windows did not know how to enumerate the drives, so it couldn't recreate the Space.
To further complicate, no-one has really figured out how to recover when using Microsoft's de duplication. So, I had one of the developers of ReclaiMe remoted in, she gave me the ability to export the entire array as a Disk image (.vhdx) which should work okay, since dedupe is automatically handled via it.
So I am restoring a single 12.6TB file, after the recovery it shouldn't be a big deal if I want to use DrivePool, since I will be removing all of the data from the vhdx. Which given, is an extremely unique (and hopefully one time) problem.
Still deciding if it is a good fit for me or not, I really like the ability to use parity. As I have a fair bit of data that I want to be able to sustain a failure on, but using DrivePool for this, would require me to have almost double the storage for duplication, and I don't have the ability to grow to that size at this time.
A normal RAID setup, isn't a good fit for me, as I need to be able to add drives on the fly, of mismatched size.
Storage Spaces will work okay for me, but I dislike it's calculations on how much space it makes available- it just doesn't follow a logical convention (for example, I create a simple space (no redundancy)- add all drives, and have a total capacity of 15.5TB- yet it only allows me to create a volume of 13.6TB, With 1.82 unused that I can create another volume with. Which makes no real sense.
I want the flexibility and parity of storage spaces, but with a logical, intuitive interface like DrivePool. I am managing about 27 hard drives, + two SSD's for caching, and + two more in RAID 1 for OS. All drives are now controlled by an HBA, so I don't run into this problem in the future.
Wow, this really devolved into a rant.
Thanks for the help.