Hello,
I ran KSlackcheck downloaded a whole bunch of updates yesterday (except kernel* packages).

When I'm lazy, I just run:
'upgradepkg *.tgz'
and since there were several dozen packages, I wasn't going to try doing them all by hand.

I noticed though that I was starting to get errors (I wish I'd written the text down before it scrolled off the screen, something about the package not being installed correctly) with some packages: first with the 'bin' package, then with 'glibc' packages. I stopped the upgrade process but I think it was too late.

Now I get several messages on boot about /bin/sh needing GLIBC_2.4, and GLIBC_2.4 is missing (which is strange because I think I had glibc 2.3.x previously installed and one of the upgrades I downloaded (and broke?) was for glibc 2.5.x). And then it stops booting and just leaves me with these messages.

I tried running the original setup program on my old (10.0) Slackware CD, but it kept wanting to do a FRESH install, and all I want is to remove then re-install a couple of new packages. I couldn't find a way to mount the existing installation and then install the handful of packages on to it. Anyone know how I can get around that, or maybe a better way to fix this?

Thanks.

You needed to install glibc first. It's a good idea to read the upgrade.txt. Slackware current has had a major upgrade. Anyway, this post might help.

http://www.linuxquestions.org/questi...d.php?t=543647

Thanks for that link!
I also read the upgrade.txt file (yeah yeah I probably should have done that earlier) and noticed I needed a newer kernel - I am running 2.6.15.4, and needed >=2.6.18. As it happens, I already had the source for kernel 2.6.20 (I think) but hadn't yet compiled it.

Anyway, I looked at the instructions on that page you linked (http://www.linuxquestions.org/questi...d.php?t=543647) to and was able to eventually successfully chroot into my system. On reboot though, I got a pile of
"Fatal: Kernel too old"
errors during boot, when loading sendmail, X, ALSA, and OSS (and maybe some others). Then the system just hangs.
I rebooted withe Slackware 11 recovery/install CD and tried to chroot back in, only to get more "FATAL: kernel too old" messages. So I went through all the steps on that page and chroot worked again (I've tried this a couple times just to confirm, and yes, chroot works fine and I can access my system until I reboot, then I have to upgrade the damaged packages with the install CD all over again). I set about configuring and compiling the kernel, and then I tried to run lilo and got back:
"Warning: root at /mnt has no /dev directory
fatal: cannot open: /etc/lilo.conf"

...ok so this is weird because I know with the slackware 10 recovery disc, I could mount my partition to /mnt/hd and then run lilo when windows overwrote the bootloader. I'm not sure why lilo won't run when I've mounted my partition with the slackware 11 cd to /mnt or /mnt/hd

help??

Ok, so for anyone who's interested, the problem above was solved very simply by setting "ROOT=/". The advice on fixing the broken slackware installation (http://rlworkman.net/howtos/glibc-recovery) advised setting ROOT=/mnt to trick the package tool. so I compiled and installed the new kernel, it boots nicely, etc...

and now I get a whole pile of messages that say "symbol _dl_out_of_memory, version GLIBC_PRIVATE not defined in file ld-linux.so.2 with link time reference"
or something to that effect (I'm mostly going from memory), and the system stops booting.