Adding drivers at the end of an restore.

[Terry Bates]

Hey There,

I have to say this is the third time I have used this project for deployment situations and its fantastic.

I have an odd request, I would like to add a custom some script to copy driver files from a directory on my fog server to my xp image once its restored to the machine.

I understand that you can mount the init.gz image and modify it, I am thinking that adding a bit of code to the “fog” script in /bin will do but at what point should I add this? for the most part my images are single partition.

Also any suggestions on what I should write as well would be great I can write a vb script file but first time with bash script, basically would like to copy a folder full of drivers say /mnt/drivers to ~c:\sysprep\drivers, anything basic will do I’m sure I can modify it to suit

There may even be an easier way of doing this but I dont know it

Cheers

[Terry Bates]

Had some time today, added this to my fog script

[CODE] if [ “$type” == “down” ];
then
# Windows XP
if [ “$osid” == “1” ]; then
echo “”;
echo " * Copying driver files";
echo “”;
ntfs-3g -o force,rw $part /ntfs
mkdir -p /ntfs/sysprep/drivers
cp -R -f /images/drivers/* /ntfs/sysprep/drivers
umount /ntfs;
fi
fi[/CODE]

It seems to work well, will probably expand on this in the future, might even send through page for your wiki with what I’m doing for deploying images.

[jdd49]

Here is a solution that I currently use. I added an extra host option in the database. Similar to kernel options but called it scripts. Here I can add different scripts to different hosts or groups. That information is then appended to the bootimage in a variable called hostScripts. I created a new nfs share called scripts. In here are all of the different bash or shell scripts that can be ran after a host has imaged and before it checks out. Here is the function I call from my boot image.

if [ ! “$hostScripts” = “” ]; then
runScripts
fi

function runScripts()
{
mkdir /scripts
mount -o nolock,proto=tcp “$serverIP:/scripts” /scripts 2>/dev/mntscripts
echo “Running Script(s)”
sleep 1

scriptCount=$(grep -o “,” <<<“$hostScripts” | wc -l)
scriptCount=$(( $scriptCount + 1 ))
echo “> Found $scriptCount Script(s)”
sleep 1
i=1
while [ $i -le $scriptCount ]
do
scriptArray[i]=$(echo $hostScripts | cut -d"," -f$i)
echo “> Running ${scriptArray[i]}”
sleep 1
cd /scripts
./${scriptArray[i]} $hd
i=$(( $i + 1 ))
sleep 1
done
}

So if hostA had hostScripts=script1,script2,script3 set for it. It would run all 3 scripts that are stored on the server.
I know this has a lot info missing but this is a general idea of how you could do it.[/i][/i][/i]

[De Thegeek]

Hi

I created this page some time ago

[url]http://www.fogproject.org/wiki/index.php?title=Working_device_drivers_snapins[/url]

I’m used to deploy device drivers as snapins. most device drivers have a installer able to run silently. However, some drivers need to run when a user is logged on. Some of them (sometimes realtek hdaudio or ac97 depending on the computer, or some ATI video drivers) are a bit ugly and don’t run well silently.

Have a look, that’s quite nice

Now I’m using .cmd files as snapins, and they start my device driver from a SMB share. This is easier to maintain and troubleshoot. No more need to create a sfx with 7z