On Sun, 2003-11-09 at 00:49, [EMAIL PROTECTED] wrote:
> I've had problems with my computer for a week or so now, and have been trying to
> diagnose the exact problem. I'm not exactly sure what is the matter, but if I
> relate the symptoms I hope someone will be able to assist me.
>
> Initially I assume the fact that problems started when I began using Mandrake
> 9.2 is purely a coincidence.
>
> Several weeks ago I noticed problems with applications randomly crashing, most
> notably those that ran most often, eg nautilus. Sometimes even my whole system
> would crash and require reboot. After a reboot I would notice that some files
> were just missing (I assume corrupt). Most notably these were icon files, as I
> would have desktop items without icons, but also certain application files would
> be missing/corrupt.
> As my main hard drive was several years old and had been getting noisier I put
> it down to a bad hd and replaced it.
> However even with the new drive I was getting these crashes, and they seemed to
> occur more often. And when a reboot was unavoidable I would reach lilo, but then
> get a message that a CRC check failed and the system was halted. This would
> happen several times in a row, and after leaving the machine off for a while, it
> would happily start and give me several hours before it all went to hell again.
>
> I chated to a few friends and they suggested that perhaps my IDE cables were
> getting old/damaged and needed replacing. Being the cheapest thing to replace I
> did so, and this has not helped.
>
> Sometimes I am logged into a terminal when it 'fails', and I see some sort of
> buffer journal error (can't recall exact words) however sometimes I can still
> type in commands, but am returned an I/O error and no programs will run - and I
> can't open a new connection to my machine, I can't open a new virtual terminal
> in screen, can't su to root, or even ls the directory; however last time it
> happened I could cd between directories, and btlaunchmanycurses was running in a
> virtual terminal and kept going, with the screen updating even.
>
> Another possible source of errors was my RAM (3 sticks of 256MB pc133 sdram). I
> installed memtest86 and tested my ram in several different patterns (all in,
> only 1 in, 2 in, mixing around slots). memtest86 returned only 1 error when all
> 3 sticks were in - noting the error to be at 318MB - as this is more than 256,
> and less than 512 I thought it was a good guess my middle stick of ram was at
> fault, and pulled it. Tests on the last 2 sticks usually revealed no errors..
> but now and again did.
> So I borrowed a stick of working 64MB pc100 sdram from work (my mobo supports
> this) and ran my system for several hours.
> I worked the system hard, my swap was on the go consistantly, and I was
> switching between applications, playing video files, and had firebird and
> thunderbird open just to keep things active. While I received no immediate
> errors I decided that I would run fsck on my partitions.
>
> I switched runlevel into single mode, and ran fsck.reiserfs on a partition on my
> brand new hard drive, and it failed with:
>
> hda: dma_intr: status=0x51 {DriveReady SeekComplete Error}
> hda: dma_intr: status=0x40 {UncorrectableError} , LBAsect=28151643, sector=6217496
>
> I then ran fsck.ext3 -c /dev/hdb1 (this is a new-ish drive that has never seen
> problems).
>
> During the check it gave me:
>
> hdb: dma_intr: status=0x51 {DriveReady SeekComplete Error}
> hdb: dma_intr: status=0x84 {DriveStatusError BadCRC}
>
> This CRC error reminds me of the CRC error when it would fail to boot up. So I
> am thinking that maybe my ram is not at fault, nor my hard drives, but that my
> motherboard is.
>
> Perhaps the IDE channels are dying, or the ram slots, or some other part of the
> mobo that I just don't know enough to be able to understand the errors.
>
> I don't know how to really test my motherboard, and I have pretty much got to
> the limit of my ability to test for problems.
>
> Can anyone offer me any advice of how to test for mobo problems? Do my errors
> sound like anything you recognise?
>
> Oh.. and no, I don't have an LG cdrom drive ;). Just a pioneer dvdrom and a
> plextor cdrw.
>
I think you've dome a pretty good job of diagnosis here, but it's one of
those classic hardware problems that require swapping of major
components to really pin down. What's left is your mobo as you've
suggested and your power supply. A faulty PS can cause this sort of
behavior, so I would be tempted to try swapping that first, just because
it's easy (and cheap if you have to buy one).
I have a feeling in my bones that you are going to be in the market for
a new mobo however. Perfect excuse to upgrade that CPU while you're at
it! ;-)
good luck!
Brian
Want to buy your Pack or Services from MandrakeSoft?
Go to http://www.mandrakestore.com
