Interesting problem with 3ware RAID and HDD failure

I have a Windows Home Server box with a 3Ware 9500S running 4x 250 GB Samsung Spinpoint drives in a 750 GB array. The RAID software reports that the RAID is degraded (specifically Drive 0) but drives 1-3 are okay. If I replace drive 0 and initiate a rebuild, the rebuild fails and an error pops up with Drive 2. A rescan of the drives makes the error go away. This seems rather odd to me.

I’m backing everything up right now, but I’m wondering how to proceed. If I had money, I’d simply buy new drives, but I’m skint.

Paging drachillix, engineer_comp_geek, and the other usual suspects.

Any thoughts, guys?

Have you tried it from the card BIOS? Has the new drive been in a 3ware array previously?

Yes and no respectively. Also, if I initiate a rebuild with the original Drive 0 in place, I get the same error with Drive 2

What is the error? Did you get a verified full backup done?

3DM2 = the 3Ware control software - reports that there was an error with Disk 2. I now have a backup, but I’m concerned: is the fault with Disk 0 or with Disk 2? I don’t want to make a new array with a faulty disk.

On a 9650SE (granted, a bit newer than your 9500), a single cable error can mark the drive as faulted (and, if you have an I2C front panel connection, light the red fault light). Doing a full verify of the drive with the 3Ware utility will clear that fault. On the other hand, errors like “pending uncorrectable sectors” will not trigger the fault light, so it is a good idea to run something like smartmontools to keep you informed.

Of course, you’ll want to back up your data first, since an error on a 2nd drive in a RAID 5 array will result in loss of data.

You could pull the drives and test them individually with something like drivesitter.

That’s 404-compliant for me, but since I’ve got everything backed up, I’m going to try WHS v2.