I'm using SuSE 10.1 and I was trying to upgrade glibc-2.5 to 2.7. So I went ahead and did "rpm -e --nodeps `rpm -qa | grep glibc-2.5`", which is probably a pretty bad idea. Now I can't execute anything. RPM won't work, ls won't work, find won't work. Any command I type would give me "bash: /usr/bin/XX: No such file or directory". However the binaries still exist when I open up konqueror to view them, I just can't run them. I need help on this ASAP since I do my work on this computer. Thanks for your time.

Removing glibc will pretty much render the machine completely useless. The only way you are going to be able to get it back up and running is to boot the machine to a Live CD and reinstall the package that way, since you won't be able to install it from the running system.

To that end, this should probably be moved to the SuSE forum, since the method to get you back up and running is going to be specific to that distribution.

I was told that I could find more help here, so here goes...

Is there a way to reinstall glibc? I have the SuSE 10.1 DVD with me right now and am wondering if there is a way to fix this.

I can't answer the question on how to re-install glibc.

I can suggest that in future, it you want to upgrade an application, instead of removing one version, then installing the new version (with rpm -e and rpm -i), you should consider using rpm -U <packagename> (upgrade an existing installation), if the application will upgrade that way. Most will; a few may not.

This is an extremely bad idea, lesson learned

You can try installing the glibc packages by using the static version of rpm if this is installed;

rpmi -ivh glibc-<the rest>

Use the locate command to find if this is available; locate rpmi

/usr/lib/rpm/rpmi is usual location of the command mentioned in prior post.

I've moved this to the openSUSE/Novell forum where hopefully uou can find a detailed solution to your problem.

The two threads have now been merged

Big lesson learned . I was able to cd into the directory jlightner mentioned, but was not able to execute anything. The only thing working are bash shell commands.

You can try using rpm2cpio and cpio also, hopefully this will work, example;

rpm2cpio < the_name_of_the_rpm_package_here > glibc.cpio
cpio -i -d < glibc.cpio

Copy the resulting contents (the files and directories) to the appropriate locations on the root filesystem (/).

1 members found this post helpful.

As noted before, nothing works beside shell commands. The only programs that works are the ones already loaded into memory.

I don't want to restart my computer just yet, because SuSE wouldn't likely to be able to boot up. However, I do have a barebone version of SuSE 10.1 installed in another partition and *might* be able access the filesystem on the main partition. I am wondering if I would be able to copy the files for glibc from the barebone version to the full version so as to fix the problem. I also have the install DVD. Would that help with anything?

I've tried googling for such a situation, and haven't found much about this situation. So, I wonder if:

1) boot the barebones 10.1
2) create a mount point for the problem partition
3) mount the problem partition
4) download the glibc rpm from the Suse repository (the same one you deleted)
5) move or copy that rpm to the problem partition
6) chroot into the partition
7) try to run rpm -i from there

I confess I am grasping at straws.

1 members found this post helpful.

Thanks for all the help, guys! I was able to boot into failsafe mode and mount the problem partition as bigrigdriver suggested. I then proceeded to curl the rpm package and followed the suggestion by Lenard.
Now everything is working normally as before. Once again, thank you all very much!