How does RIS 'know' which image to download?

[Rob Savage]

I have a test environment of one desktop and one laptop, plus the RIS
server.

Both have been imaged multiple times during testing - the desktop with
XP and the laptop with W2K.

I have just riprep-ed both for the latest image of both OSes and then
started both up simultaneously to download the new image to ensure
everything is working OK.

Here's the weird part - the laptop has downloaded the W2K image while
the desktop has taken the XP image.

That's exactly what I want of course, but how did they 'know' which
image to take?! I have automated all the .OSC menu screens so that RIS
simply installs the top choice in the OS selection menu. So why would
would that have different results on these two machines?

My thought is some sort of record of the machine's MAC address being
linked to the image, but wouldn't that interfere with imaging the same
machine using a different RIS image if you needed to?

Any ideas?

Cheers

Rob

 

[NIC Student]

Some images may not be available when you choose a RIS image. This can be
caused by file permissions on the *.sif file in each image or becasue the
machines use different hals.

I greatly prefer to use custom setup and file permissions on the sif files
to regulate RIS imaging rather than customizing your osc to grab one image.
Custom setup is set in group policy so you can allow it for one set of users
and not the others if you want.

 

[Rob Savage]

Scott

Thanks for the reply.

The HAL thing evidently answers this - my dell dekstops and Toshiba
laptops obviously differ in their HAL and so they only see the
relevant images. (KB 289638 covers this).

This dead good actualy, as I can now publish both my images and be
confident that the right machines will pick up the right image!

I am looking to minimise the user's role in setup as much as possible,
and currently this is to the extent that they do not login via the OSC
screens (we've hardcoded a RIS admin account & password), so I don't
think the custom setup would apply in our case.

However, as we build up a library of images, that may well be
something we come back to!

Cheers

Rob

 

[Guest]

We've used 3 different HAL's until recently (Hyperthread is #4). You can find the HAl type in one of the files created after pushing the image up on the server.

We broke it down like this:
PC - this is a non power managed system. The message "It's now safe to shut off your computer" is usually this HAL.

ACPI: Normally a laptop, occassionally some desktops.

ACPI 1.0 UP: Uniprocessor desktop

ACPI 1.0 MP: Hyperthread or server. Although we did just see a PC MP server come through...

4 Images are hard to maintain if you have alot of clients; and want to keep them patched, configured, and up to date with the latest applications.

Good Luck!