This originally started as a post on the Endpoint Management forums and as a cry for help, but halfway through I figured out the solution on my own. I decided to go ahead and post it for anyone else who might be running into this problem, and it was suggested that I submit it as a help article. I hope that this article will save others the hours of frustration and self-doubt I went through before stumbling across the solution.
PROBLEM:
I'm working on a managed delivery policy to deliver MS Office 2007, the "Save as XPS or PDF" plugin and the "Office 2007 Help Tab". I actually have a mostly functional policy in place with some test users, but due to some issues with the Help Tab I wanted to test adding a reboot parameter into the end of the policy run.
So I cloned the existing policy, renamed it with a "TESTING - " prefix, made the changes I wanted and saved it. I built a fresh PC from image, placed it into our Testing OU in AD, installed the Altiris client and all the plugins and then assigned the test policy to the PC.
Two policies actually applied to this machine as I have another that is enterprise wide already in place. That policy, a basic script that runs and cleans out temp files from the PC, ran without incident as it has on every other PC in my enterprise. The "testing" office 2007 policy is only HALF running.
Basically, the policy loads up and runs through it's detection and applicability checks for each of the 3 tasks in the policy. Those work exactly as expected on a PC without the software the policy is designed to install. Then it just stops. When you open the tasks from the agent window, they show the package as "Disabled".
Now, this policy works just fine on the other PCs it is already assigned to. I can go right on my own PC and open the link to the package on the site server it is pulling from. If I uninstall Office from my own PC it reinstalls right away. so I know the package is good, and both my machine and the test machine are using the same Site Server to pull the package from. They are also the same OS, the same patch level and the same hardware. Just to eliminate those as culprit.
SOLUTION:
Apparently ONE of the applicability checks was actually NOT passing. This is more of an issue than I thought it might be. When I built each of the packages, I set the applicability rules differently. For Office 2007, I set the applicability to "Windows XP Professional - English".. This is what we use on our desktops currently, and I didn't want to accidentally install this on a server OS. So that made sense, and works just fine. HOWEVER, I recently changed the Applicability check for the "Office 2007 Help Tab" to check for the presence of Office 2007. THAT one was failing.
So why should this matter? I mean, it makes sense that the applicability check should fail for that task, Office isn't installed yet. Once Office installs then the task to install the Help Tab should run, right? NOPE. Apparently, if you chain several tasks together in a policy, none of the tasks can be dependent on a previous task in the chain. If even one of the applicability checks fails, the whole policy fails.
So, for relatively new people like me, please be aware: Check your Applicability on each task you are putting into a Policy. Do not make them dependent on each other or you will have a broken policy on your hands.