Xorg corrupt after a failed install from Rawhide (How to recover)

[Jim Cornette]

Williams Jr, Ernest L. wrote:
> Hi,
> 
> I tried to install "xorg-x11" via up2date from rawhide a few days ago on my FC3T1 desktop.
> The retrieval wnet okay.  As a matter of fact, I tried to update a lot of things including the kernel and many other packages.  Maybe, I tried to do too much at once.
> 
> Anyway during the installation phase xorg-X11 installation failed; therefore up2date aborted before completing everything.
> 
> How does one recover?
> 
> I removed the xorg-xll packages.
> 
> Now I try:
> =================================================
> [root at matrix root]# yum install 'xorg-x11*'
> Gathering header information file(s) from server(s)
> Server: Fedora Core 2 - Development Tree
> ===================================================
> 
> However, yum appears to be hung since nothing is happening
> 
> 
> Thanks,
> Ernesto
> 
> 

To view the log without a GUI, you might run
cat /var/log/Xorg.0.log |less
and scroll through the file to see why X departed on you.

I had a problem where X was halted in mid-update before with a previous 
test release. I retrieved the latest X versioned rpms from the repo 
(using mc, of course) and installed it with rpm locally. I believe it 
was rpm -Uvh *.rpm using the downloaded X files (before xorg-x11 and 
xfree split). I cannot recall if I needed --nodeps or not. If you 
uninstalled w/ nodeps you might need to.
In my case, there was a perticular font package used by X that did not 
take when upgrading X. This caused an abrupt halting of the install and 
the adding and removing of programs never finished.
There were two different versions of the font package in the directory 
on the mirror repository used.

Now after all the runaround views. It might be easier to run
rpm -qa |grep xorg
to see what versions of X you have and to determine if you have two 
different versions of the same package installed. If you have version-4, 
version-8 and other possible combinations. This could be messy. If they 
are all the same versions, then you are alright.

*or*

You can log in as root. Then cd to the /var/spool/up2date directory. 
 From there, type ls *.rpm and see which rpms downloaded before the 
crash. The rpm downloaded with the latest date is probably the reason 
for the crash and probably is corrupted (not completely downloaded)

Anyway, try running the below command in the /var/spool/up2date directory.
rpm -Fvh *.rpm
This should give you an error message related to the damaged rpm downloaded.
rm -rf whatever-damaged.rpm and confirm delete.
Try above rpm command above and most likely, it will complete.
If you did delete your X packages. It might be needed to run.
rpm -Uvh *.rpm within the directory instead of freshen.

I hope you get the basic concept from this attempted explanation. I'm 
getting lost myself. It is easier done than said.

Jim