I recently found a major bug with the udpated glibc packages on Red Hat's updates.redhat.com FTP server for Red Hat 9.0. If you try to upgrade to the new glibc packages, the process may fail leaving your server in very bad shape. Any command that you run will return "Segmentation Fault", so nothing will work including logging in. Your server will be unbootable when you restart and will hang at something that says "init: no more processes left in this runlevel", because the root filesystem won't be able to mount.
I searched the web and it seems that I'm not the only one to run into this problem. If this happens to you, here's how I got the server to boot again. You will have more problems to deal with afterwards, but getting the server back up and running is most important.
Boot off of the RH90 CD #1 or a boot.iso CD. Enter rescue mode by typing "linux rescue" at the boot: prompt. When it asks to mount the root filesytem, say Yes.
After your at the shell, find out which glibc packages are installed...
rpm -qa --root /mnt/sysimage | grep glibc
... then delete them all individually using something like this....
rpm -e glibc* --root /mnt/sysimage --nodeps
Next, lets re-install the glibc RPMs from the RH90 CD...
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-2.3.2-5.i386.rpm --nodeps
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-common-2.3.2-5.i386.rpm --nodeps
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-devel-2.3.2-5.i386.rpm --nodeps
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-kernheaders-2.4-8.10.i386.rpm --nodeps
Type exit to reboot from rescue mode. At this point, your server should be able to boot again.
I turned this post into an FAQ here -->
ChrisP
RHCE, LPIC-1, CCNA, CNE, MCSE, +10 others
I searched the web and it seems that I'm not the only one to run into this problem. If this happens to you, here's how I got the server to boot again. You will have more problems to deal with afterwards, but getting the server back up and running is most important.
Boot off of the RH90 CD #1 or a boot.iso CD. Enter rescue mode by typing "linux rescue" at the boot: prompt. When it asks to mount the root filesytem, say Yes.
After your at the shell, find out which glibc packages are installed...
rpm -qa --root /mnt/sysimage | grep glibc
... then delete them all individually using something like this....
rpm -e glibc* --root /mnt/sysimage --nodeps
Next, lets re-install the glibc RPMs from the RH90 CD...
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-2.3.2-5.i386.rpm --nodeps
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-common-2.3.2-5.i386.rpm --nodeps
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-devel-2.3.2-5.i386.rpm --nodeps
rpm -Uvh /mnt/sources/RedHat/RPMS/glibc-kernheaders-2.4-8.10.i386.rpm --nodeps
Type exit to reboot from rescue mode. At this point, your server should be able to boot again.
I turned this post into an FAQ here -->
ChrisP
RHCE, LPIC-1, CCNA, CNE, MCSE, +10 others