There I was minding my own business when out of the blue, I could
not log in. The exact sysptoms were about 1 second after the login screen
appeared, the screen would blank, there would be periodic very brief flashes
onscreen, and periodic disk activity. Ctrl/alt/del would bring back the login
screen and shutdown activity as normal. No faults shown in /var/log/messages,
or on /var/log/boot.log. e2fsck gave a clean bill of health to the linux drives.
The system is a K6/2-500, 64MB ram, Appollo MVP3 (?) chipset, Award Bios, 8MB S3
Trio agp card, 2.2.14-15 kernel, and disks: hda = 2.5 gig; hdb1 = /boot; hdb2 =
swap; hdb3 (4 gig) = /; hdc =cdrom. All are ide, which is built into the
kernel. The version is the version 6 gold pack, with the 7.0 update.
So I installed fresh on a new hd, 6.5 gig put in as hdb, set up the same as
the first one. (/dev/hdb1 = /boot; hdb2 swap; hdb3 /) This was a repeat of
the previous agony. Every "install" cd I got from Mandrake was dodgy; there is
no sensible way to recover from a dud rpm in the system, and the install takes
it upon itself to shut down, in case you could do something. It installed make
without gcc, missed i586-glibc20-linux/lib files needed by Netscape. I tried
for them, and found the rpm was broken:-(. This is an opportune moment for
a special word of thanks to whoever in mandrake left PPP support out of the
2.2.13-7 kernel. Prayers were said for you here, sunshine!! But I complain,
therefore I got running ;-). I reinstalled my dodgy system as /dev/hdd, even
putting it in lilo. This is not strictly kosher, as ide 1 has a master(cdrom)
which may be unseen to some bios routines, and mebbe some hardware.
When I check the old installation, it seems fine; when I e2fsck the disks,
they come up fine; when I boot the old installation (on hdd) it locks booting on
hard drive optimizations for hdd; when I boot the new installation with
/dev/hdd1 & /dev/hdd3 in /etc/fstab, it does the same, but it can 'optimize'
/dev/hdd if it's not in fstab :-0. I can then mount it on /mnt, and /mnt /boot
and can access it. I did a couple of wildcarded 'cp' commands, and they would
crash out after a while Once it spat at me that it had disabled the dma on
ide1. When the 'cp *' commands would crash, I'd get ' file not
found.........ide 1 reset : success' and then I'd get the prompt back. When you
use a '*' there's NO FILE NAMED ! Why is it saying file not found :-/ ? I
thought only windoze did that :-(
Can anyone tell me
1. what caused the first system to die 1 second after login prompt? There's
very little running on it.(atd, cron, gpm, network, linuxconf, lpd, kudzu, and
local - none of the network protocol stuff which could have delays in there).
Last thing showing in /var/log was linuxconf. I killed that (Interactive
startup) but it didn't help. I wouldn't claim to be fully competent with linux,
but I can access this old install if such is needed.
2. Are the hdd problems the result of it's unkoshered positioning, or is the
disk the problem? Why does e2fsck pass it when the system can't optimize it?
The disk in question is a Seagate ST34311A, and about 1 year old.
--
Regards,
Declan Moriarty
Applied Researches - Ireland's Foremost Electronic Hardware Genius
A Slightly Serious(TM) Company
Don't you get browned off with stupid quotes on the bottom of sig files?
Keep in touch with http://mandrakeforum.com:
Subscribe the "[EMAIL PROTECTED]" mailing list.
