Jeff Spaleta wrote:
On 11/14/05, Jim Cornette <fct-cornette@xxxxxxxxxxxxxx> wrote:
Anyway, many programs had broken scriptlets. The list of problem
packages was quite long. I ran rpm -Uvh <package> --noscripts on the
broken scriptlet packages and am aware that this probably negatively
impacted the correct installation for these packages. Of course using
-ivh --noscripts on the kernel was useless. :-)
when booted into the 1656 kernel, i had some interesting problems
trying to use mkinitrd, even manually. I still don't know why that
happened, but the problem with mkinitrd when booted into 1656 was the
underlying problem trying to install new kernels. As soon as I booted
into an older kernel the problems with mkinitrd went away and I was
able to install kernel rpms without any problem. I am now booted into
1663 and able to install kernels via yum and rpm without issue.
I have not had a problem with any other scriplets. The other failures
your are seeing are most likely a cascade effect of from having a a
large update transaction set fail part way through... leaving a trail
of broken deps.
-jef
November 10 to present is where the kernels downloaded but never
installed. The deps are met. broken deps are caused by openssl being
upgraded and later packages to fail installation because of the
scriptlet errors while the installation is underway.
The trouble is that the package downloads but does not install. Yum is
happy to install other packages where the deps are met.
It leaves the system in a bad state. Post and pre script error leave db
entries for removed rpms and pre end up having packages downloaded again
and again for the same package that should have installed on th last run
of the yum program.
A fresh install is in order when test one is released. (In my case anyway.)
Jim
--
Whenever you find that you are on the side of the majority, it is time
to reform.
-- Mark Twain
--
fedora-test-list mailing list
fedora-test-list@xxxxxxxxxx
To unsubscribe:
https://www.redhat.com/mailman/listinfo/fedora-test-list