[CentOS] CentOS 3.8 Kernel Update with NVIDIA Video Card

Bisbal, Prentice PBisbal at LexPharma.com
Mon Oct 23 12:11:57 UTC 2006


-----Original Message-----
From: centos-bounces at centos.org [mailto:centos-bounces at centos.org] On
Behalf Of Alfred von Campe
Sent: Friday, October 20, 2006 10:43 PM
To: CentOS mailing list
Subject: Re: [CentOS] CentOS 3.8 Kernel Update with NVIDIA Video Card

On Oct 20, 2006, at 22:22, Corwin Burgess wrote:

> I can think of two ways to solve this problem but I'd rather have some

> expert advice. What's the best way to boot with the new kernel, 
> install the nvidia driver and of course update the NVIDIA kernel 
> module?

The following worked for me:

   1. Download the latest driver packager from NVIDIA (e.g., NVIDIA-
   2. Boot the new kernel to run level 3 (or wait for it to fail to go
to run level 5)
   3. Log in as root and run ./NVIDIA-Linux-x86-1.0-8774-pkg1.run and
answer the prompts

Oh yeah, you will most likely need to have the kernel-devel RPM
installed so that it can rebuild the kernel module for CentOS.


You don't need to reinstall the lated NVIDIA driver every time you
update your kernel. In fact, this will cause problems if you need to
revert back to your old kernel. You just need to install a new kernel
module compiled for that version of your kernel. If you install the
entire driver package, you will be installing new versions of all the
related libraries, too. Then if you revert to the earlier kernel, that
kernel module be an earlier version that doesn't match the version of
the newer libraries, and you'll have a similar problem. Then to fix this
one, you'll have to install the earlier kernel sources to recompile the
kernel module for that kernel... And so on, and so on... 

If you know you'll never revert to the earlier kernel, it's okay to just
install the whole new driver package. If you might revert, it's better
to learn how to install just the kernel module. I forget the exact
syntax, but if you do 'man nvidia-installer', all the details are there.
Even better, HP supplies the nvidia drivers as RPMS. These RPMS are for
RHEL, and include a script /etc/init.d/nvconfig. At startup, this script
checks to make sure that the current kernel has an nvidia module. If it
doesn't find one, it installs just the kernel module, no fuss no muss.
The HP RPMS are just the nvidia drivers repackaged into RPM form. I
recommend either using those RPMs, or at least extracting the
/etc/init.d/nvconfig script from the RPM and be done with it. That's
what I did. 


The contents of this communication, including any attachments, may be confidential, privileged or otherwise protected from disclosure.  They are intended solely for the use of the individual or entity to whom they are addressed.  If you are not the intended recipient, please do not read, copy, use or disclose the contents of this communication.  Please notify the sender immediately and delete the communication in its entirety.

More information about the CentOS mailing list