ontap will only ever use a disk that is at lest as big as the failed disk. In other words, ontap may: grab a 144 to replace a 36 but it will not grab a 36 to replace a 144

Also, since you have FC9 shelves, you may have 36GB Zoned checksum disks. If they are, then you can replace them with either zoned or block checksum disks. However, at the 36GB capacity, zoned disks are 34.5G and block are 34G, therefore, if your filesystem is block-checksum based, then you can only use a block-checksum disk. If your filesystem is zone-checksum based, then it can be replaced with another 34.5G zone checksum disk or any larger disk.

--tmac

Patwardhan, Nate wrote:

All,

I am migrating a 740 so that the shelf/volume configuration looks like the following:

vol0 -> 1 x DS9 @ 36g (raid4) data -> 1 x DS14 @ 144g (raid4)

Basically, I wanted to not waste a 144g drive on vol0 while maximizing my space on the DS14. Note that we're also going to be consolidating many of our NetApps in the future, so it made sense (to me, at least) to keep 'data' on a separate shelf.

However, I have concerns with the above configuration. I'm not concerned with disk failures on 'vol0' (or any other volume that might live on 36g drives) since I'm pretty sure that OnTap will do the right thing and take a 36g drive from the spares pool. I am pretty concerned that 'data' might pull a 36g drive from the spares pool instead of a 144g drive.

Is my concern justified? Will OnTap 6.3.x do the right thing for the 'data' volume and pull a 144g drive from the spares pool or is there the chance that it would pull a 36g drive? Or is OnTap smart enough to know that a 144g disk with more than 36g utilization should be replaced with a disk of at least 36g?

-- Nathan Patwardhan, UNIX System[s] Administrator, Sr. Akamai Technologies, 839B/8CC x 83035 npatward@akamai.com AIM: notoriousnvp71