[colug-432] bulk/automated updating machines with NVidi drivers

[Greg Sidelinger]

I solved a similar problem with the vmware vmxnet driver with an init
script that ran before the network script. It would try to load the
module and if it failed it ran the vmware-configure script. Below are
the guts of the init script from a sles 10 box. Since you are
compiling from source you should be able to tell the vmware module to
compile with out any prompting. Another other solution is to package
your own driver RPM that gets installed during your yum updates.


# chkconfig: 2345 9  0
# description: Run vmware-config-tools.pl if needed
#
### BEGIN INIT INFO
# Provides: vmware-config-tools
# Required-Start:
# Should-Start:
# X-Start-Before: $network vmware-tools VMware_Tools autofs
# Required-Stop:
# Should-Stop:
# Default-Start:    2 3 4 5
# Default-Stop:      0 6
# Short-Description: Run vmware-config-tools.pl if needed
# Description:       Runs vmware-config-tools.pl with the "default"
#       option if the vmxnet module can not be loaded.
### END INIT INFO


SCRIPT="/usr/bin/vmware-config-tools.pl"
test -x $SCRIPT || { echo "$SCRIPT not installed";
        if [ "$1" = "stop" ]; then exit 0;
        else exit 5; fi; }

SCRIPT_OPTIONS="default"

case "$1" in
    start)
        echo -e "VMWare Config Tools: Checking to see if vmware tools are setup"
        modprobe vmxnet
        if [ $? -eq 0 ]; then
                echo -e "VMWare Config Tools: Look to be setup right"
        else
                echo "VMWare Config Tools: Running $SCRIPT"
                $SCRIPT $SCRIPT_OPTIONS
        fi

        # Remember status and be verbose
        rc_status
        ;;
    stop)
        echo -e "VMWare Config Tools"

        # Remember status and be verbose
        rc_status
        ;;
    restart)
        ## Stop the service and regardless of whether it was
        ## running or not, start it again.
        $0 stop
        $0 start

        # Remember status and be quiet
        rc_status
        ;;
    *)
        echo "Usage: $0 {start|stop|restart}"
        exit 1
        ;;
esac
rc_exit







On Fri, Dec 11, 2009 at 1:35 PM, Scott Merrill <skippy at skippy.net> wrote:
> Hi everyone.
>
> I've asked a couple times about administering groups of like machines.
> One of these days I'll actually start using Puppet, or a similar
> administration automation tool, but that's not likely to be today. :(
>
> We have a lab with half a dozen Linux workstations, all running RHEL
> 5. We have a number of engineering and modeling applications, for
> which 3D acceleration is a requirement. I've got the NVidia binary
> drivers installed on these computers.
>
> Every time a new kernel package is released by Red Hat, I'm forced to
> manually run the NVidia installer again. I'd like to automate at least
> that little piece right now.
>
> How are others tackling this? I know the NVidia installer has
> command-line options: is the recommended process to extract the
> installer and script the installation?
>
> Thanks!
> Scott
> _______________________________________________
> colug-432 mailing list
> colug-432 at colug.net
> http://lists.colug.net/mailman/listinfo/colug-432
>



-- 
It is a mistake to think you can solve any major problems just with potatoes.
  - Douglas Adams