Auto Apply Drivers only as a last-ditch effort?

[Jaybone]

I was thinking that since we have a good number of random systems (5 of X, a dozen of Y, 8 of Z, etc) out in the wild that we don't have driver packages for, but that mostly have well-known hardware at least for the NICs (plain ol' Intel or Broadcom stuff), it might be useful to have an Auto Apply Drivers step in a task sequence, and set it to only run if none of the Apply Driver Package steps were triggered.

What I'm not seeing, though, is any real built-in way to easily test this. The best thing I can think of is to set up a step that deliberately fails before all of the Apply Driver Package steps. My reasoning being:

1. The deliberately failing step should set _SMSTSLastActionSucceeded to false
2. If none of the driver packages are applied, _SMSTSLastActionSucceeded stays false
3. If a driver package does get applied, _SMSTSLastActionSucceeded gets set to true (hopefully, heh)
4. Auto Apply step checks the value of _SMSTSLastActionSucceeded, and only runs if false

Seems kind of kludgy to me, though. Is there a better way to do this?

[TrialandError]

I completely understand where you are coming from. I took over an environment that only used Auto Apply drivers and didn’t even use categories to limit them. I implemented driver packages but had to have a way to let the lesser used hardware still be able to utilize the Auto Apply step. I will walkthrough what I ended up doing but you are correct, there is no good way to do this. The way I did it was a little convoluted but it worked well for what we needed. The reason it worked so well is that I could exclude one model at a time instead of having to create all my driver packages at once.

 

I created a new group in the task sequence labeled Drivers

I then create steps to install my driver packages and have corresponding WMI queries for each step.

The last step is Auto Apply Device Drivers. I use the inverse of all the WMI queries that I used on the driver packages as a condition of this step.

 

 

 

 

Hope this helps.

[Jaybone]

Didn't even think of that, but that seems like a much cleaner way to do it than forcing a deliberate failure. Would need to be updated every time new packages are added, but for how often that happens... might be worth it.