On 09/28/2009 11:57 AM, Philip Heron wrote:
Kevin J. Cummings wrote:
> I'd look into the "errors" that happen way before hal tries to start.
> I'd probably force an fsck on the partition that holds /etc because it
> looks like your /etc/passwd file is messed up.... Maybe all of your
> filesystems.... Then, again, perhaps this is the result of running fsck
> on your filesystems after the power failures. You may be SOL if you
> were running fsck and took a power hit. If that's the case, a
> re-install may be your best bet.
I suspect this might be the case - the second and third cut where only a
minute apart, the machine would have been in the middle of starting up
again.
> Do you have file fragments in your top
> level lost+found directories?
e2fsck -f says it's clean, lost+found is empty. I tried starting up in
runlevel 2 and although it boots up a lot quicker there are still a load
of errors (mostly concerning missing users or UIDs) and I can't log in.
The most I can do is restart it with Ctrl+Alt+Del.
You want to boot into runlevel 1 (single user mode). If you do it
right, you'll already be the root user. If it continually prompts you
for a password, you'll need a "rescue" disk to boot from.
I don't know what all files are involved in the user database,
but
passwd and shadow seem fine? root and my own user are in there.
Attack your errors one at a time, starting with the first error. Check
the /var/log/dmesg and /var/log/boot.log files for more information.
-Phil
--
Kevin J. Cummings
kjchome(a)rcn.com
cummings(a)kjchome.homeip.net
cummings(a)kjc386.framingham.ma.us
Registered Linux User #1232 (
http://counter.li.org)