Douglas A. Tutty wrote:
On Sun, Mar 09, 2008 at 03:59:45PM +0000, postid wrote:
For the second time in a month I got an error message
indicating bad inodes and had to fsck manually.
I've had bad inodes before not long after a failure to load
my PCMCIA modem (which resulted in endless error messages on
boot) and before I knew alt sysrq-r -s -e -i -u -b and had
shut down by shutting down the power. I've had the modem problem once
about a week ago and rebooted with RSEIUB.
Oh yes, and about a week ago I was trying a USB stick and
the machine froze and I did an RSEIUB.
This time the machine froze in KDE, so I did RSEIUB. Upon
rebooting, I got "Inodes that were part of a corrupted orphan linked
list found" and got the suggestion to run fsck manually.
That resulted in "Inode 820174 was part of the orphan inode
link. Fix?" and a few saying the same for some other inodes.
It then asked permission to fix a deleted inode with zero
dtime, clear an unattached zero-length inode, fix inode bitmap
differences, and fix a wrong inodes count for group #50.
Any ideas as to what causes such problems? Is it that PCMCIA
loading problem, my hard drive dying, an OS problem or
a software problem?
The magic keystrokes just sync the disks, they do not unmount the
filesystems. Thus, things can become corrupted. If it were me, after
such a reboot, I'd come up in init=/bin/sh and run fsck manually.
Please pardon my ignorance here, but what do you mean by "come up in
Ideally, you'd use ext3 with data=journal. That way, syncing the disks
will get the data _somewhere_ on the disk so that replaying the journal
in a normal boot fsck would set things right.
Again, my apologies, this time for not supplying more complete info. I'm
using ext3, running Sarge on an IBM R40 laptop along with Knoppix (hd
install) and WinXP (for encrypted DVDs.
Well, ideally, the system should be stable enough not to need a reboot.
If this is Etch, check for bug reports.
If the hard drive is dying, there should be some errors in
/var/log/syslog. Also, install smartmontools so that you can check the
S.M.A.R.T. data on the drives. If smart tells you that the drive is
failing then believe it. If it tells you that everything is fine, take
it with a grain of salt. Always have good backups.