F11 on T400 will not boot after failed resume from hibernate

[jbkt23]

I did a hard reset after the resume stalled. When I rebooted there were errors and fsck automatically cleaned up. On the subsequent restart the boot hangs after entering interactive mode with the phrase:

/usr/bin/rhgb client: error while loading shared libraries: libm.so.6: cannot open shared object file: No such file or directory

it continues to load a couple other items in the init list and repeats this line three times before it times out.

the last line printed to the screen is:

init: rcS post-stop process (1237) terminated with status 127

This is not kernel specific as I have tried booting into the two previous ones with the same result.

I do have a backup of this system but I have never tried to restore a whole system. I dual boot or I should say tri-boot between F10 and vista on this laptop and I am able to boot into those systems. I can mount the F11 partition in F10.

Any suggestions on solving the specific problem?

[jbkt23]

Well, stumped the stars I guess. What I have found is that /usr/bin/rhgb-client is called from rc.sysinit so there is no way to bypass it by selecting a lower run level. I also found that libm.so.6 is a link to libm-2.0. It and its target exist in two locations under /lib and /lib/x86_64/noseng/ ...

I don't have the machine in front of me so the second path is from memory. When I get back to the laptop I will see if the links are broken and go from there.

[jpollard]

Originally Posted by jbkt23

Well, stumped the stars I guess. What I have found is that /usr/bin/rhgb-client is called from rc.sysinit so there is no way to bypass it by selecting a lower run level. I also found that libm.so.6 is a link to libm-2.0. It and its target exist in two locations under /lib and /lib/x86_64/noseng/ ...

I don't have the machine in front of me so the second path is from memory. When I get back to the laptop I will see if the links are broken and go from there.

You might look in the lost+found directory of your root volume - It just might
have the library, though the name '#<nnnnnnn>' will make it hard to identify.

Along with a lot of other files...

Happened to me due to a power failure during an update... I guess I was
lucky - none of the libraries lost prevented the system from coming up.

I ended up doing a forced yum install of all packages to get everything back.. even though it took several days.

[jbkt23]

I've found the missing file and replaced it and am now back in F11. The target file (libm-2.10.so) in /lib64/ for libm.so.6 was missing. So after booting into F10 I did:

rpm -q --whatprovides libm.so.6

the answer was glibc-2.9-1.i686.rpm.

I then did a search of the rpmpkg log for glibc in the F11 partition

less /oldrt/var/log/rpmpkgs | grep glibc
glibc-2.10.1-5.i686.rpm
glibc-2.10.1-5.x86_64.rpm
glibc-common-2.10.1-5.x86_64.rpm
glibc-devel-2.10.1-5.x86_64.rpm
glibc-headers-2.10.1-5.x86_64.rpm

I downloaded these rpm's and then used archive-manager to browse them and find the matching file I needed.

What is the best way to query all the installed rpm's to see what files are missing?
I just want to see the error/missing files listed in a text file that I can browse. The first part is to use rpms own facility but I want to pipe the errors only to a file.

[jpollard]

Good question.

The only way I could figure out (didn't implement it though) was to go through the installed packages, extract
a list of all files installed (package-name, file) and put them into a data file. Unfortunately, I fear this step
could take a long time to create.

Then go through the data file and check that each file existed... Recording those that didn't exist.
The output would be a list of packages that needed re-installation, Since that would likely have
duplicate entries it would need to be passed through a "sort -u" to get only the unique package
names...

A final "yum reinstall `cat file`" could then try to reinstall them, unless you know/trust the
rpms you MIGHT have kept around after installing them the first time...