Help. I'm dead in the water. Bad file number for /dev/ttyx
Ack, I can't boot, and I can't figure out the problem. Any
help is appreciated; even if you don't want to read to the end
of the message, at least some hints on where to find docs will
help (no, I can't read linux man pages).
Symptom is that when I boot, I get something like (I had to write it down):
INIT: version 2.71 booting
INIT: Entering runlevel: 2
Debian GNU Linux 1.3 (none) tty1
(none) login:
Note that none of my boot scripts were run... When I log in as root,
with the correct password, I get:
<some date> login[8] unable to change tty '/dev/tty1' for user 'root'
Unable to change tty /dev/tty1: Bad file number
I can't figure out what's going wrong. If I Ctrl-Alt-Del, it won't
reboot, but instead asks me for a root password like it's trying to
go to single-user mode (this may be a debian feature, I don't know).
I have to push the reset button in order to reboot.
If I try to boot with 'single' as a parameter, again it asks for a
root password (some security feature?), but when I give the correct
password, it says "Entering runlevel: 2", rather than going into
single user mode. (my guess is that it tried, but then the shell
failed, and it assumed it was time to go to multi user mode).
Finally, using the rescue disk, I changed inittab to have tty6 run:
bash -i < /dev/tty6 > /dev/tty6 2>&1 But then when I rebooted, it
said something like respawning too fast, disabled for 5 minutes.
That's odd, because that sort of command worked just fine with one
rescue disk using 'ash'. If there's some sort of way to properly have
a shell run without using a rescue disk, I'd like to know how (so I
can test various commands, devices, etc; using a rescue disk doesn't
tell me anything, as it boots just fine)
Fsck found some problems but fixed them. I've tried rebooting with
older kernels too.
Any ideas on which binaries may be bad: init, getty, login?
Any idea where to ftp these, without them being in a .deb
package? Would my libc be bad?
(This all started when I was having trouble with dselect. I had
installed a downlevel version of 'dpkg', and was futzing around with
dselect; I used the 'unstable' variety, so I could use 2.1.x kernels,
but had not had any problems, and had rebooted many times since using
last using dselect. Today, dselect couldn't find the 'unstable'
directory, and when I checked manually, I noticed the new 1.3.1.r3
thingy, and ftp'd and installed dpkg from there, as the readme said.
I did not touch the libc stuff. Noticing that it actually downgraded
dpkg, I tried to get the old version back. Then dselect started
having core dumps, which were traced to the dpkg-ftp 'setup' package,
which croaked on line 137 (or 135?) when trying to open the ftp
connection. The *only* packages I had touched were perl, perl-debug,
dpkg, and dpkg-ftp. In frustration, I rebooted, but it never came
back up.)
--
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: