Re: Very weird behaviour - hardware failure ?
Updated reflection:
I am beginning to wonder if all this psychadelic stuff isn't
related to the initial install and physical and logical disk geometry.
My first experience with Linux was Slackware. I got nothing but a suite
of hda: errors when I tried to make the ext2 fs checking for bad blocks.
When it sais it was finished I could never get any further until I
finally realized that it was using the logical geometry from the bios
that DOS needs. It was not doing any translation to the actual physical
geometry. I had to tell the install process, at the install boot prompt,
about my actual geometry.
When i installed Debian, I noticed that it too was reporting hda
paremeters from the bios record during the boot process so I quit and
restarted. At the new installation boot prompt I reported the physical
geometry of the HD. The boot messages still reported the logical
geometry, and I went with that, not knowing what else to do.
Is this OK ? Does Debian handle making translations between
logical and physical geometry ? If not, how do I report the physical
geomtry to Debian ?
Thanks again everyone,
Gerald
On Thu, 16 Oct 1997, I wrote:
> My system is behaving very strangely since I installed Debian. I had
> Slackware on my box before and it seemed to behave as one would expect. I
> was despondent at first because it seemed that I could not configure
> Debian to behave as I come to expect with Slackware. Now I suspect
> something far more insidious and hope someone out there will recognize the
> problem. In short, am I experiencing a hardware failure ?
>
> The symptoms involve hda errors, kernel panics, making a ppp connection
> with PAP, and freaky xdm light shows. Personally I am beginning to
> suspect that my hard disk has gone, but don't know enough to say for sure.
> Maybe something else is going that makes it look like the HD has gone.
> Maybe its software somewhere.
>
> Symptom # 1 Preparing my Linux partitions for the Deb install.
>
> When initializing an ext2 Linux native partition with a bad-block
> scan things started out fine with
>
> checking for bad blocks (read only test): xxxxxx/542776
>
> but then things started to get nasty and I got messages like the following
> interspersed with the above message
>
> hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error ]
> hda: read_intr: error=0x40 { UncorrectableError }, LBAsect = xxxxxxxxx,
> sector=yyyyyyy
> end_request: I/O error, dev 03:07, sector yyyyyyy
>
> I got about a gazillion of these error messages. At first I thought that
> all of a sudden something had happened to my disk that was causing blocks
> to go bad, but I am no longer so sure. I did get the install finished and
> Debian seemed to be working.
>
> Symptom # 2 Disappearing DOS partitions
>
> Before I installed Linux I had DOS primary partition and an extended
> partition containing two logical drives. When I bought the computer, I
> knew I would be installing Linux, so I only used about a third of the disk
> for these DOS partitions. I didn't have to use fips or anything similar
> to get ready for Linux.
>
> After installing Linux everything seemed hunky-dory. I could access C:,
> D:, and E: from both Linux and form DOS or Windoze 3.1. Then one day, in
> windoze I went to File Manager to look for something on D:, and neither D:
> nor E: was visible -- no little driver icon to click on on the driver icon
> bar.
>
> I can still see them from Linux however.
>
> Symptom # 3 Seeming random hd i/o timeouts
>
> I can't remember the exact message to this error and can not find where I
> wrote it. But it was hard disk i/o timeouts that started with "hda:
> status error" or somesuch, then said "hda: drive not ready for command"
> then something or other about resetting and things being okay. This never
> caused a crash or anything but happened doing things like an "ls". I
> could get these messages in the middle of a directory listing if the
> directory was long.
>
> I haven't had this problem for awhile.
>
> Sympton # 4 Kernel Panics
>
> I foolishly didn't write down all of these and now can't remember what I
> was doing when these panics happened. They were always during ordinary
> things though. The most recent happened when I was trying to get my new
> ppp connection setup with pppd and chat. I was just testing a new script
> and whammo! Dirty powerdown.
>
> Just found a message from one of the first times this happened
>
> message from syslogd
> Kernel: Kernel panic: EXT2 fs panic (device 03:07): ext2_read_inode: unable
> to read i-node block - inode = xxxxxxxx, block= yyyyyyyyy
>
> When I tried to shutdown -r now I got
>
> hda: status error: status=0x59 { DriveReady SeekComplete DataRequest Error ]
> hda: status error: error=0x40 { UncorrectableError }, LBAsect = xxxxxxxxx,
> sector=yyyyyyy
> end_request: I/O error, dev 03:07, sector yyyyyyy
>
> got that four times then
>
> hda: drive not ready for command
> bash: /sbin/shutdown: I/O error
>
> Symptom # 5 pppd not working
>
> I'm not sure if this one is me or my new ISP (not the freenet) but I will
> mention it anyway. pppd connects, I get assigned my IP address, ifconfig
> reports the ppp0 interface and the lo interface, and route -n reports the
> host, the local loopback and the gateway. Then I can do nothing. Can't
> telnet, ftp or browse known sights, can't even ping the host I am
> connected to and it can not ping me (talked to my ISP). No one else seems
> to having problems. The server was happily working with other dial-in
> clients while we spoke and the ISP was able to ping them. This is a
> connection that was working.
>
> Then as I mentioned in the above symptom, a kernel panic was provoked.
>
> Symptom # 6 xdm light show
>
> The latest in this series of frustrating little phenomena is the strobe
> effect I was getting from xdm. From xdm I did a control-alt-F1 to get
> back to the console. I was going to shutdown but then wanted to go back
> to xdm to do something. I hit alt-F7 as I have down in the past. The
> screen did the flicker thing it does when xdm is starting, then flickered
> back to the console then to xdm and I was caught in a handy little cycle
> that I could not stop (cleanly). All the while I could hear my hd arm or
> whatever makes those little grinding sounds humming in rhythm with the
> flicker of the screen. I tried control-C and just giving the shutdown
> command anyway.
>
> Had to hit the restart button and do an fsck. When I went to start xdm
> from a fresh boot after a clean shutdown, ie. xdm was the first thing I
> typed after login, the flickering started and I could not start xdm nor do
> anything else.
>
>
> Please help. I now have the fsck and its switches memorized. This is a
> command I would rather be forced to look up when needed than to have
> commited to memory.
>
> What is wrong ? These things have been happening with increasing
> frequency and increasing diversity. I have this sinking feeling in the
> pit of my wallet that it is my only slightly more than a year old hd.
>
> Any thoughts ?
>
> Thanks in advance,
>
> Gerald
> ye416@freenet.victoria.bc.ca
>
>
> --
> TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
> debian-user-request@lists.debian.org .
> Trouble? e-mail to templin@bucknell.edu .
>
>
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-user-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: