rawhide report: 20060714 changes

[Tom Brinkman]

On 16 12:55:S, Jim Cornette wrote:
> Tom Brinkman wrote:
> > On 16 10:56:S, Mike Chambers wrote:
> >> On Sun, 2006-07-16 at 10:12 -0500, Tom Brinkman wrote:
> >>>      I was hopin you or someone else might know ;)
> >>
> >> I acutally ended up just using the mouse middle button.  I
> >> just typed in yum remove and kept on highlighting and middle
> >> clicking and doing that few at a time until they were gone.
> >>
> >> --
> >> Mike Chambers
> >> Madisonville, KY
> >
> >      OK, but it's still a PITA ;) And we still don't know what
> > caused the issue, so there's a good chance it could happen
> > again
>
> Killing yum or rpm before they complete the cleanup phase will
> leave database entries in the rpm database. I have killed either
> and had to do a cleanup afterwards.
> What worked for me was to go to the cache where the rpms were
> cached. I knew I was bad, so I did not run yum afterwards. :-)
> Anyway, from the cach directory with the yum packages, move any
> rpms like the kernel out of the way. Then you want to run:
> rpm -Uvh --replacepkgs --replacefiles *.rpm
> on the cached rpms. This will clean the older packages out of the
> database and leave only the latest version installed by
> overwriting the files once again. Any bad rpms like zip for
> instance will exit, but others should be one instance per
> program.
>
> Probably the cause was mkinitrd and the kernel locking up for
> some reason and railing nash to peg the cpu. Killing nash during
> the update and not killing yum or pup should work.
>
> Alternative way,
> Jim

    Thanks Jim, I believe you're right. FWIW, the last 3 kernels had 
the 'nash' problem. So for the last 2 I --exclude=kernel\* an do it 
only after the other updates are successful. It was probly that 
first 'nash' kernel that messed up the db, as I believe I did kill 
yum during that one.

    Thanks again, an I'll try an keep this in mind
-- 
    Tom Brinkman                   Corpus Christi, Texas