Monday, June 11, 2007

Hard Drive Woes

And this past weekend one of my hard drives went berserk. At first I didn't know what was wrong as I just upgraded Ubuntu. The linux kernel was also upgraded, and after reboot it halted at disk identification. For some reason the 80G Western Digital disk caused problems. Some DMA timeouts and it all seemingly locked up. Well, only seemingly, as I found out later, the kernel needs about 3 minutes to give up on it. I found this unacceptable, so I disconnected the disk and rebooted. And now I had a ghost disk in the same place, albeit 300G Seagate, giving me the same error. It just somehow duplicated there. Okay, I switched off the onboard controller in the BIOS on which the now disconnected disk was hooked on. The same error yet again. Only, now I also had a ghost controller. G'damn, I go boot Knoppix, and everything is well. Hm! I go boot the previous Ubuntu kernel, and everything is well. Bloody hell, they messed up something, again. Last year they did it with XFree -- I couldn't work for a day in graphics mode. This time they did it with kernel. Only now there's no forum overflowing with spits over Ubuntu; it seems only I'm affected. Well, I really don't have the time for this. They patched something around IRQs, and I bet this is where my errors come from.
And the story continues... I mounted the disk into another computer and ran a SMART long selftest, just for the heck of it. The disk went working and after 10 minutes it switched itself off and went playing with the busy light. Oh I really wanted to just throw it out the balcony. My divine W800JB Western Digital disk, with 8MB cache and what not, from 2002. Bloody hell. I switched off and on the machine 5 times and the disk finally seemed to work. I wanted to check the SMART selftest log and the disk again spun down (as in switched off) and started the lightshow. 5 another power cycles and there it is, seemingly working. I wondered if this need for power cycling 5 times was arising out of the fact that I may have interrupted the SMART selftest, because according to standards drives should resume with selftest if interrupted. So I issued the SMART command for cancelling the selftest. Again the drive spun down, only now there was no lightshow, now the busy light was on constantly. And no 5, 10 or 20 power cycles helped, so I turned the computer off and went for some denervousiation.
And this morning, well, I couldn't help it but turn the computer on and see what's with the disk. Wow, it seemingly worked. Oh, no more SMART shit! I checked the filesystem, it passed. It's readable. So I turned it off and went to work. I'll turn it on again when I'll have a new disk onto which to copy the contents of the old one. And then I'll boot it out the window.

2 Comments:

At Friday, June 22, 2007 3:54:00 AM, Blogger Dav said...

Ahahaha not so SMART huh?
And the new font is a bit hard to read on a LCD TFT man, maybe change it or get it larger?

 
At Sunday, June 24, 2007 6:26:00 PM, Blogger Pastel River said...

Postanje enkrat na leto je too little too late.
Prekmurc, če ne bi blo tega bloga se midva ne bi spoznala. Tvoja naloga je, da vsaj približno brca. :-*

 

Post a Comment

<< Home