Daniel B. Thurman wrote:
Since 'yum update' was executed in a messed up selinux state, I am not
certain that all of the updates was correctly performed for all of the
files updated as some files were deposited/installed and yet post-installs
may have failed as well as evidenced with the gnome/kde #prelink# issue
noted above preventing me from logging into the console as a non-root
user. I will search for all the #prelink# files but it is impossible to
catch other things that may have been missed.
Anyone know how I can force-reinstall all the newly downloaded rpms or perhaps
force install all of the rpm's in the database which presumably has the updates
as well?
Dan
You are correct that failures with the post scripts will leave
incomplete installation of packages. I heard on the test list that pam,
which is pretty important for logging in was effected by the post script
problem.
rpm has two options to overwrite your currently installed packages
--replacepkgs and --replacefiles
You could try these options on your locally downloaded rpms. I guess
these should do the replacement of all content with a corrected content.
My favorite method of new is to erase the database entry with the
--justdb option and then run yum install messed-up-package. Yum does not
know it is installed and tramples over whatever is in its way. (I assume)
Depending on how frequently you clear your yum cached rpms, you could
use the versions within the var/cache/yum/<my-repo> directory.
for rpms with pre script failures, they download again and again, but
never install. Running rpm -Fvh from the yum cached rpm directory
should install the rpms which failed on the pre script routines.
Jim
--
Don't shoot until you're sure you both aren't on the same side.
