May 11th, 2005

Tall ships (porthole)

Not so SMART numbers

Certain numbers to do with the SMART status of the system disk here have just got worse (and it's done that "wedge the IDE bus for a bit while it thinks about stuff" trick a couple more times. I'm getting a few errors and warnings from atapi and disk in the event log (and those are ones that you really don't want to see). Hmm.

Ironically, a disk I recently added was in fact one of the two spare disks my dad had in case one died (he held on to the larger one).

Update: Windows just presented me with this:
Delayed Write Failed : Windows was unable to save all the data for the file \Device\HarddiskVolume1\Program Files\WindowsUpdate\wuaudnld.tmp\cabs\com_microsoft.894320_W2K_SP5_x86_WinSE_154269\Windows2000-KB894320-x86-ENU.EXE. The data has been lost. This error may be caused by a failure of your computer hardware or network connection. Please try to save this file elsewhere.

I am not amused. Now where did I leave my DriveImage disks...
  • Current Music
    God Put A Smile Upon Your Face ~ Coldplay/A Rush Of Blood To The Head
Tall ships (porthole)

It's not good when winlogon.exe goes bye-bye

Well, I'm now on the laptop as shortly after that last post Bad Things happened. Lots of bad yellow paging errors appearing in the System log, lots of evil write failures (I had recently turned the write cache back on, not that it made any difference here), the shell died, and then apparently winlogon.exe died because the system bluescreened on me. It's never a good sign when you get write failures in directories and in such files as $MFT (master file table, the NTFS equivalent of the FAT table).

So, my plan now is to image the drive (or at least the bits that matter), then swap it out before it goes and dies on me.
Tall ships (porthole)

Update on the disk:

Update on the disk:

Officially, according to the IBM/Hitachi disk fixer-type-thing, it works fine. SMART is happy as anything, disk surface is good, it's sweet as pie.

Unofficially, according to Windows 2000 and DriveImage, the disk is screwed. It's hard to argue when you get delayed write failures on such files as $MFT, INACCESSIBLE_BOOT_DEVICE bluescreens, random reboots, and read errors.
Tall ships (porthole)

Day 1 of Computer Dissection

Lessons learnt so far in the dissecting of this computer:

Knoppix doesn't seem to like it if I unplug my hard disks and then boot from the second cd drive.
My main cd drive is either dying as well (probably through a dislike of being run on its side, though it was happy enough when knoppix was working) or does not like it when i unplug my hard disks.
The Promise IDE controller takes forever to do disk detection if there are in fact no disks (or, more accurately, an unpowered disk).
Oh, and the IBM either has a very mangled file system, or its dying (and the disk utility for it refuses to admit that).

Of course, I could have a dead stick of memory.

Current plan of action is to install Win2k to the Maxtor drive on the Promise controller, then run a brute-force xcopy (runas LocalSystem, copy everything, continue on errors) to the Maxtor drive, doing all of this with the data drive disconnected. Then run a low-level format on the failing disk and see if that makes it behave.

If anyone knows a way to mount either a disk or a NTFS partition read-only under Win2k, then I'd be very grateful. Hopefully that way all I'll need to worry about will be read errors.
  • Current Music
    Faithless album on hi-fi