Questions Re: downgrading glibc to 2.3.2 on RedHat Enterprise
Hello Linux folks,
I've run into a serious problem with the C runtime libraries on my linux Redhat Enterprise Server 3 box; I have experience with UNIX but am fairly new to Linux, and have some basic questions I need help with to fully understand what's happenning.
I downgraded to version 2.2.3-5 of glibc. (The reason being, I was following a recipe for installing Oracle 8 on Linux). Unfortunately, while doing this, I got a 'script error' message from rpm, and immediately after the OS was pretty hosed; doing an 'ls' command resulted in Floating Point Exception. I don't remember the exact format of the rpm error message, but is that scenario familiar to anyone here ? Also, I cannot recall if I used the --oldpackage flag when installing glibc-2.2.3-5; does anyone know if not doing so would have serious repurcussions ?
The kernel cannot load at boot time either; when I reboot I get a load of messages that look like this:
/bin/sh: relocation error: /lib/i686/libc.so.6: symbol _dl_catch_error,
version GLIBC_PRIVATE not defined in file ld-linux.so.2 with link time reference
INIT; LD "x" respawning too fast: disabled for 5 minutes
I've used the repair disk to look at /lib, and I think it may be possible to fix the problem there, but I don't know enough about the C runtime on RedHat to know what libraries to create links to. (Anyone know what a 'relocation error' is? My guess is it's something to do with loading libraries into the kernel. Also, what is "link time reference" ? My guess is that symbol is undefined in the ld-linux.so.2 lirary. Unfortunately, I don't have access to nm in repair mode.)
I'm thinking my best course of action now is to reinstall Linux and leave the parititions unchanged, but, does anyone know if and how I could fix this problem with a repair disk, and poking around /lib ?
Also, does rpm leave a log file anywhere ?
Thanks,
Colm.
|