Seriously messed up system - is there anyway back?

[Jim Cornette]

Paul F. Johnson wrote:
> Hi,
> 
> 
>>If you still had the yum cache, you could drop to runlevel 1 with 
>>selinux=0 and change to the yum development cache. Make sure there are 
>>no kernel rpms in the cache and run rpm -Uvh with maybe --replacepkgs 
>>and --replacefiles as options. For kernel rpms, the rpm -ivh should work.
> 
> 
> As you can see, this happily worked. I wonder what went fertang?
> 
> TTFN
> 
> Paul

SELinux in my case (now fixed in policy). It downloaded and gav the 
impression that the packages were installed. Querying the "just 
installed" packages showed no package at all installed for particular 
packages queried.

Your case might be related to the same temporary SELinux problem or 
something along those lines.

As suggested by another reply, you ought to query your installed 
packages and verify that you do not have multiple entries with older 
package rpm versions in the database along with the latest rpm version.

Also querying all packages as root for missing files might be a good 
decision. (chroot and firewall rpms show missing files as a regular 
user, they check out alright as root). Expect the missing files for 
mozilla and for cups. All other files should not have problems. (in my 
case anyway).


If you do get missing files other than mozilla or cups, you can get and 
reinstall the package again. I used rpm -e --justdb and followed it by 
yum install <package-with-missing-files>. I believe there were a couple 
of multiversion rpms and a few missing files before reviewing the 
package contents via rpm queries.

Jim