February 08, 2005

Pictures of My Beagle for Shannon

The whole reason I built a new PC was that my old system, The Unimog, suffered a cascade of drive failures.

The whole reason I built a new PC was that my old system, The Unimog, suffered a cascade of drive failures.

I started out with:


A directory on the #C singleton started acting seriously flaky. As I was trying to salvage its data to my #A RAID one of the elements of the mirror was announced to have failed. Good grief. So I bought a Western Digital 80GB drive (WD800JB SE), swapped out the apparently-broken element, and told the RAID to rebuild itself. 10% into the rebuild cycle, the remaining element died. I managed to suck off most of the user data onto a 300GB maxtor firewire drive and then I threw away the computer.

Now that I have Unimog's successor, Kamaz, working, I started sifting through the data I had rescued from Unimog. I was horrified to discover that I had somehow overlooked my 30GB music/ directory! Then I remembered a suggestion Stas made -- load the drives on an external USB enclosure and see what data can be recovered.
These USB enclosures turn out to be really simplistic -- a cheap case, a power cable, a usb cable, and a small IDE -> USB circuit . I can't even find a brand name on the box! But who cares, the things work. (*) I went through each drive to see what data I could recover.

I loaded up one element of the #A raid. There were a few directories in My Photos/ that took a long time to copy data from. I could hear the drive making a lot of strange high-pitched straining noises. I had to abandon one large photoshop file in one directory, but I think I eventually got everything else.

The #B Maxtor drive was fine. It never failed while it was on Unimog. I was shocked how incredibly loud it was compared to the Western Digitals. It had a piercing, shrill, high-pitched whine that sounded as if I was runnign a turbine under my desk. Although it's apparently sound, I can't imagine using this drive anywhere near me -- the noise is terrible. In contrast, the Western Digitals are essentially silent.

The #C Maxtor is apparently totally dead. I plug it in. I hear it spin up. I hear the PC give the "USB appliance plugged in" noise, I see nothing, and then I hear the "USB appliance unplugged" noise. So I guess this drive is seriously smoked.

Ok so what to do?

I think it's pretty clear #C is a total write off.

The replacement WD800JB drive I bought for the #A RAID never had a chance to get tainted (*well..... unless the motherboard was frying drives with power?), so I just reformatted it, and I guess it can be a #2 external backup drive. The Western Digital Diagnostics utilities gave it a clean bill of health.

The #B Maxtor, although usable, is too loud to be bearable. If I was honest with myself, I'd just throw it in the garbage right now, but probably it will end up in a forgotten drawer somewhere, to be thrown away in a couple years

Now what to do with the original two elements of the #A raid. The one that failed first I didn't test in the USB enclosure, maybe it's cooked, I don't know. But what about the other one, which I was able to get almost all the data off of, but there were some patches that were very slow to read? I guess the question is "is there fundamental physical/electronic damage to the drive?" or "is the problem with the drive some corruption of the filesystem that a reformatting would solve?" The concern being, "yeah, reformat the thing, but there is still secret damage and this drive is going to recorrupt itself later." Again, I guess if I was being smart, I'd just say, "this drive has been tainted. It's a tiny and comparatively old drive anyway. It should be immediately tossed in the trash." I guess I'll run the diagnostics on it anyway, just for the hell of it.

Ok, I just finished the diagnostics. During the first run it said it detected bad sectors "that might be fixable." So I told it to fix them. It seemed happy. Then to double-check, I re-ran the 45 minute test suite. It ran fine and said the drive passed. It's not really clear to me what a 'bad sector' is anyway, so I am not sure what my comfort level should be with this drive. Second-string back up? Scratch disk?

I might as well run this on the other Western Digital element of the raid, the element that originally 'failed' and see if this drive can be salvaged too.

I should run the Maxtor Diagnostics on the #B Drive.




* well, works to a point. This morning as I was trying to remove a hard drive from the enclosure I managed to shear the pin-head from the ribbon cable, ruining it. Fortunately it looks like I replace the entire (2" long) eide cable. I'll make sure the replacement has a small ribbon handle on one end.

Posted by Nils Blutig at February 8, 2005 04:51 PM | TrackBack