cat /proc/mdstat gives me: md1 : active raid5 sdi1[4](S) sdh1[2](F) sdg1[1] sdf1[0] 5860145664 blocks super 1.2 level 5, 512k chunk, algorithm 2 [4/2] [UU__]

The history of this array is that it was a functioning RAID5 array, that wasn't really being used. I pulled one of the disks out for other uses temporarily, then wiped it and returned it, the idea being that it was returned to the array as a "new" disk to replace the "failed" (removed) one, albeit that it was actually the same disk.

This is the RAID array that's been causing problems via a USB caddy (I now have a cheap SATA card that seems to work reliably).

I was able to mount the array (albeit degraded because it was missing a disk) with the disk removed. Now I can't (makes sense, 2-out-of-4 disks in a RAID5 array isn't going to work very well!).

My reading of mdstat is that sdi1 is currently "spare" (I think that's the one that came out and went back in), and that sdh1 has failed. Seems like quite a coincidence.

As I said, the data on the array wasn't important, I'm using this as an opportunity to learn more about what's going on because next time the data may well matter.

Mark