|
Home > Archive > Red Hat Installation > February 2004 > "/ contains a filesystem with errors" and other fun things
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
"/ contains a filesystem with errors" and other fun things
|
|
| felice 2004-02-05, 2:35 pm |
| I've finally gotten round to trying to get Redhat 7.3 going on my machine
again, with very mixed results. The current problem is this:
After a normal shutdown, I got the warning "/ contains a filesystem with
errors, check forced" then "Error reading block 2686981 (Attempt to read
block from filesystem resulted in short read) while doing inode scan" on
starting up. Running "fsck /" as instructed returned a number of errors
such as "entry has deleted/unused node" and "inode X ref count is Y, should
be Z". Restarting after this, I get the text login screen flashing up several
times, but X doesn't start. Running "startx" reports "failed to initialise
the NVIDIA kernel module!". I could try reinstalling the NVIDIA driver, but
I'm reluctant to redo anything before fixing the source of the filesystem
corruption. Do I need to replace the hard disk, or could there be some
other explanation?
History:
Up till now I've been using Windows NT 4.0 sp6; I'd been intending to
switch to linux for a while, but had trouble getting the dual boot working.
I recently had some spare time, and finally managed to get GRUB to load
Windows (previously I'd just set the BIOS to boot straight from the Windows
drive), though it wouldn't load linux, which reported "kernel panic: No
init found". I can't recall whether this was a new problem or the state I'd
left it in after my previous experiments. I then reinstalled linux, and got
things more or less working, but a couple of days later it stopped working
again for no apparent reason (I don't think I changed anything before the
previous time I'd shut down the computer). The error was "hdb: read_intr:
status = 0x59 {DriveReady SeekComplete DataRequest Error}" follwed by
"error = 0x40", "mount: error 22 mounting ext3". And a couple of days after
that, Grub stopped working, reporting "Error 17", even though I had made no
attempt to do anything outside Windows.
Meanwhile, Windows had started rebooting at random (a problem I'd
previously encountered while trying to rip music to MP3 off my 40x CD-ROM,
but not during normal usage), so I took my box in to be serviced. They
replaced the motherboard, and wiped and tested but did not replace my 80GB
hard drive (the one with linux - Windows is on an old 2GB drive which while
slow has been very reliable). The reboot problem appears to be gone, but
I haven't had Windows running since, and didn't have linux running
beforehand, so can't be certain that the problem was in the hardware rather
than in Windows.
Since the change to the motherboard broke my Windows installation, I decided
to focus on getting linux going instead, and reinstalled it once more. All
went well for a while, and I got sound, video, and internet working. However,
after I edited XF86config-4 to increase the colour depth to 24, then next
time I started up the Gnome Panel crashed (a segmentation fault, I think?)
and on following startups it remained missing. Running "panel" from a command
window starts it successfully, though reporting some errors, but I'd rather
not have to do that every time I log in. Nothing else reacted adversely to
the increased colour depth, so I suspect the timing of the Gnome failure was
a coincidence.
Most recently, my Firebird browser was displaying fonts badly, and a web
search suggested trying the "Linux GTK2 + XFT" version instead. While trying
to get that working, the exisitng "Linux GTK" version locked up (probably
foolish trying to run both at once). So, I shut down the computer, which is
where the problem I described at the start of this post kicked in. It doesn't
seem likely that Firebird would cause this sort of problem, but I'm no expert.
At this point my natural inclination is to try passing the monitor through
an unopened window, but since I'm fairly sure the problem is in the main
computer case rather than the monitor, I don't think I should trust my
instincts in this case. Help?
felice
| |
|
| On 5 Feb 2004 19:39:42 -0800,
felicehl@randomstatic.net (felice) wrote:
> The error was "hdb: read_intr:
> status = 0x59 {DriveReady SeekComplete DataRequest Error}"
That's indicative of a broken hard drive. It is possible to get some
errors, like that, because the data on the drive is stuffed, rather than
the drive itself, but running a test on the drive is the way to find
out.
Look for a tool for testing your drive on its manufacturer's website.
--
My "from" address is totally fake. The reply-to address is real, but
may be only temporary. Reply to usenet postings in the same place as
you read the message you're replying to.
This message was sent without a virus, please delete some files yourself.
| |
| felice 2004-02-07, 3:34 pm |
| Tim <Tim@mail.localhost> skribis:
> That's indicative of a broken hard drive. It is possible to get some
> errors, like that, because the data on the drive is stuffed, rather than
> the drive itself, but running a test on the drive is the way to find
> out.
> Look for a tool for testing your drive on its manufacturer's website.
I downloaded the seagate tool, and it reports no errors on a complete
surface scan 8( Any suggestions on why I'm having so much trouble if the
drive itself isn't faulty?
felice
| |
|
| Tim <Tim@mail.localhost> wrote:
felicehl@randomstatic.net (felice) wrote:
[color=blue]
> I downloaded the seagate tool, and it reports no errors on a complete
> surface scan 8( Any suggestions on why I'm having so much trouble if the
> drive itself isn't faulty?
I've played with a Seagate tool at a friend's place, letting the drive
scan itself found errors, getting the computer to scan the drive didn't
(both are options of the program).
If the drive isn't faulty, it could be the filing system the computers
using with the drive. Running fsck (file system check) should resolve
that, and fix it (if it can), or report that it can't fix it. You'll
have to read the manual for fsck, I've not used it enough to guide
anybody through the procedure.
You could have other hardware issues - like cabling issues (such
not-so-good cables that are getting noise from other sources), or
hardware that doesn't do UDMA very well (can't offer any help there,
either).
--
My "from" address is totally fake. The reply-to address is real, but
may be only temporary. Reply to usenet postings in the same place as
you read the message you're replying to.
This message was sent without a virus, please delete some files yourself.
| |
| felice 2004-02-14, 2:33 pm |
| Tim <Tim@mail.localhost> skribis
> I've played with a Seagate tool at a friend's place, letting the drive
> scan itself found errors, getting the computer to scan the drive didn't
> (both are options of the program).
Reinstalling the NVIDIA driver caused further errors (including
status=0x59), and then running the seagate tool in self-scan mode
reported one bad sector. I obtained a replacement harddrive yesterday,
and no problems so far. *crosses fingers*
felice
| |
|
| Tim <Tim@mail.localhost> skribis
felicehl@randomstatic.net (felice) wrote:
[color=blue]
> Reinstalling the NVIDIA driver caused further errors (including
> status=0x59), and then running the seagate tool in self-scan mode
> reported one bad sector. I obtained a replacement harddrive yesterday,
> and no problems so far. *crosses fingers*
Sounds like you had a drive that was marginally faulty (but would become
worse), so the problem wasn't always discovered. I'd be pointing the
finger at the drive being faulty, rather than anything else causing
problems with that drive.
--
My "from" address is totally fake. The reply-to address is real, but
may be only temporary. Reply to usenet postings in the same place as
you read the message you're replying to.
This message was sent without a virus, please delete some files yourself.
|
|
|
|
|