X-RAID 2 Operation and Downsizing

I tried another experiment and got a positive result. This one is not as interesting as the original, but it is worth letting the community know about in case this ever helps someone in the future.

I created an XRAID2 volume using wildly different sized drives (40GB, 300GB, and 500GB). XRAID, predictably, decided to create several different layers. It ended up doing the following:

1 40GB layer that ran across all three drives (RAID 5)

1 260GB layer shared between the two largest drives (RAID 1)

That decision left 200GB of unused space on the largest disk in the array. I have been repeatedly told that you cannot replace a disk with a smaller disk, but in this situation it would seem silly not to allow a smaller disk to fill in for the 1/3 empty 500GB drive.

I pulled out the 500 after the array finished bulding and hot-swapped for a 320 disk (which was 20GB larger than it needed to be to accommodate all of the data on the 500). Although I had just swapped a larger disk for a smaller disk, the ReadyNAS unit accepted it and rebuilt the array onto the smaller disk. I don't know how often this will be helpful, but it is good to know that the drive looks for the necessary capacity, rather than an exact replacement. Makes sense, but it's good to have confirmation.

---

Learning2NAS wrote:

 

 

(1) If I lose a disk in a RAID-5 and don't have an equal capacity replacement on hand, can I use a smaller disk to fill in the dead disk's place until the new disk arrives? (This would downgrade the RAID volume to a lower capacity for the interim).

No.  One exception might be if some of the disk wasn't actually in the array.  For instance if you had 3x2TB+3TB, you might be able to replace the 3 TB drive with a 2 TB drive.  I haven't tried this, so I don't know for sure what would happen.

---

 

(2) Since X-RAID 2 works by creating layers, does the answer to (1) change if my array was originally created on smaller disks, and my temporary disk is the size of the original disks that my volume used to be made of? Or do those layers disappear each time the volume expands?

 

The layers never disappear.  So if you start with 3x1TB, the array will always have a 1 TB base layer.  If you then expand to 1x1TB+2x3TB, the array will aways have a second 2 TB layer.  It doesn't change the answer to (1).

 

 

(3) I currently have disks of varying sizes in my array and the ReadyNAS is utilizing 100% of their capacity. If it can do this as they're added to the array, regardless of what order their added, why can't it downsize if there is free space to do so? Please explain this in detail, rather than saying that XRAID is meant to be expandable only (a common answer I've found to similar questions).

 

I think it is theoretically possible to downsize.  Basically you need to make sure that any file data that is in the section you want to remove is moved to a new safe area before you destroy it.   In principle, this could be done.  In practice I think it would be hard to make it reliable.  In any event, Netgear hasn't chosen to invest the amount of work needed to do it.  

If they did do it, it would almost certainly need to be done off-line (with the data not available).

Expanding is much simpler, since you know for certain that there is nothing in the newly added space.

 

 

(4) If anyone knows where there is a full technical specification breakdown for X-RAID 2, please post it. It would be helpful to have an X-RAID bible around for future reference.

 

---

I haven't seen a full specification. I think that it is better to think of it as a collection of expansion features, which are implemented in somewhat different ways on different platforms.

For instance, XRAID-2 in NAS running 4.2.x uses a combination of lvm and mdadm and the ext file system.  In NAS running 6.x.x, it uses mdadm and btrfs (and no longer uses lvm).   

So I don't think there is a single technical specification, rather it is a set of behaviors that various generations of ReadyNAS implement in somewhat different ways.