My computer. It has free space on its hard disk. There is nothing obviously wrong with it. And yet it is grindingly, drearily slow, spending enormous amounts of time 'waiting for I/O' when I do more than the very bare minimum with it. This situation is new in the past month or so. I am running Debian Linux (etch) with KDE. How do I sort this out? I am sick of waiting minutes to open a KMenu item.
no subject
Date: Wednesday, 12 March 2008 09:47 am (UTC)Is anything showing in /var/log/messages ?
no subject
Date: Wednesday, 12 March 2008 10:04 am (UTC)Mar 12 10:02:31 hikari kernel: hda: dma_intr: status=0x51 { DriveReady SeekCompl
ete Error }
Mar 12 10:02:31 hikari kernel: hda: dma_intr: error=0x40 { UncorrectableError },
LBAsect=2534921, sector=2534918
Mar 12 10:02:31 hikari kernel: ide: failed opcode was: unknown
Mar 12 10:02:31 hikari kernel: end_request: I/O error, dev hda, sector 2534918
Mar 12 10:02:36 hikari kernel: hda: dma_intr: status=0x51 { DriveReady SeekCompl
ete Error }
Mar 12 10:02:36 hikari kernel: hda: dma_intr: error=0x40 { UncorrectableError },
LBAsect=2534921, sector=2534918
Mar 12 10:02:36 hikari kernel: ide: failed opcode was: unknown
Mar 12 10:02:36 hikari kernel: end_request: I/O error, dev hda, sector 2534918
no subject
Date: Wednesday, 12 March 2008 10:13 am (UTC)no subject
Date: Wednesday, 12 March 2008 10:23 am (UTC)Gah, I am so lacking in clue. Additionally, I can't drop everything, as essay deadlines loom lijke icebergs in my immediate future. Bleh. I might try to transfer my essay-work to the laptop, and then I could let the desktop chug away writing DVDs while I get on with stuff.
I'm curious - it looks like there is only one bad sector; does this *really* mean OMG bail out now?
no subject
Date: Wednesday, 12 March 2008 10:25 am (UTC)I advised you to make a backup some time ago, and also suggested that disk problems might be at the heart of your slowness, but we didn't chase this up.
no subject
Date: Wednesday, 12 March 2008 10:29 am (UTC)no subject
Date: Wednesday, 12 March 2008 10:32 am (UTC)no subject
Date: Wednesday, 12 March 2008 10:44 am (UTC)no subject
Date: Wednesday, 12 March 2008 10:50 am (UTC)no subject
Date: Wednesday, 12 March 2008 05:45 pm (UTC)no subject
Date: Wednesday, 12 March 2008 10:32 am (UTC)In my experience: yes.
Modern disks are designed to be highly resistant to failures of any kind. At the point at which you start seeing clear signs of trouble that tends to mean the trouble in question is very bad.
no subject
Date: Wednesday, 12 March 2008 10:32 am (UTC)no subject
Date: Wednesday, 12 March 2008 10:41 am (UTC)As others have said, bad sector problems tend to only get worse, not better.
If time were no issue a sensible recovery plan might be "remove disks; fit new disk; install fresh etch; add old disks; copy stuff off them; remove and destroy dying old disk". As things stand, I'd suggest it'd be safest to use hikari as little as humanly possible until you can find the time to sort it out.
[*] Light heartedly, I suppose you could say that the chugging and delays are because the data missed the IDE bus and had to wait for the next one ;-)
no subject
Date: Wednesday, 12 March 2008 10:44 am (UTC)no subject
Date: Wednesday, 12 March 2008 10:51 am (UTC)no subject
Date: Wednesday, 12 March 2008 11:43 pm (UTC)What I do for emergency backups is copy crucial things to a USB thumb drive. You can get 4GB ones for 20 quid or so now, which should be enough to cover important documents.
no subject
Date: Thursday, 13 March 2008 12:10 am (UTC)no subject
Date: Wednesday, 12 March 2008 11:07 am (UTC)In my experience a small number of bad sectors will render important parts of your drive unreadable before the damage worsens. At that stage it's probably possible to retrieve the majority of your data, but it's more awkward as the disk may no longer be fully accessible and some vital data may be in the bad sector itself.
Do not do further work on the machine with a faulty disk apart from backing up - the more you use it, the more likely it will critically fail. Also, back up in strict order of decreasing importance.
The worst case scenario is that you get mechanical failure, the disk head crashes onto the platter and all your data is destroyed beyond hope of recovery other than by specialist firms. Switch the computer off and on as few times as possible (ideally : zero) before completing your backup.
On the bright side, bringing the disk up to Manchester shouldn't impact on its reliability. Modern disks are far more shock resistant than in prior years.
no subject
Date: Wednesday, 12 March 2008 03:44 pm (UTC)However, I've also seen this because of a bad drive cable, which on a desktop is super easy to fix. There are ways to check but I am saying nothing until you've backed things up :P
You should probably mention the cable idea to Dave as it's one of those too obvious things that is easy to overlook. Certainly I dug around for hours before noticing "Hmm, this cable has a bit of a serious kink in it".
no subject
Date: Wednesday, 12 March 2008 04:52 pm (UTC)*nods*
It could also be a problem with a controller card, or even a driver for that controller card.
On board controller card failure would be somewhat annoying.
no subject
Date: Thursday, 13 March 2008 10:18 am (UTC)