Hi,I'm new to Task Sequence and was hoping someone could point me in the direction of a document that points out when to use Software Distribution vs. Task Sequences. Is there a "best practice" for this sort of thing?Thanks in advance.

A) Software Distribution also cannot survive rebootsB) The only "shortcoming" of task sequences to my knowledge is that there can be no user interaction A) This is not true, if you link software distribution tasks together using the "run this program first" option, the software distribution can survive reboots, assuming you also configure the software program correctly. There are several options regarding how the program interacts with reboots, you can tell SCCM that the program will force a reboot on the machine and it will wait patiently until the reboot is complete before beginning the next step. That said, this method can be a little tedious to maintain, just pointing out that it does work.B) Just a little clarification here. You can run a script/HTA that requires user interaction during a TS. However, that script/HTA generally runs in the context of the SYSTEM account on the machine. This means that the currently logged in user is not going to see the interaction because it is not running as that user, it's running as system. The only time you can see this interaction is while the TS is in WinPE mode as WinPE is also running as SYSTEM and will therefore display the script/HTA. You can also use the "run as user" option, however the user would have to match the currently logged in user, which is probably not exactly practical for this purpose.As far as the difference, generally, Task Sequences are used for OS deployments. I don't imagine that most SCCM shops use them for much more than this, but I have seen a TS used for software deployment only when the number of steps needed to complete the software configuration was extreme (several dozen steps, which are all updated individually).Essentially it's like this: Software deployment is perfectly fine for 1 or 2, or even a few steps that need to be done, but much beyond that and you should use a TS as it's just much easier to work with.Scott Gill SCCM Consultant

A is true though. A single program executed via software distribution cannot survive a reboot -- the reboot terminates that program. The scenario you are describing is multiple chained programs via software distributions. Thus, if you were to create a script to perform multiple tasks and one of those tasks, in the middle of the script is to reboot, the script would not continue running after the reboot. Perhaps more of a semantic difference and ultimately both are correct depending on the configuration, but A is definitely true.Jason | http://myitforum.com/cs2/blogs/jsandys | http://blogs.catapultsystems.com/jsandys/default.aspx | Twitter @JasonSandys

Depending on how the script is written, A is still not true... For example, if the package is interupted SCCM will report that the program failed because an unexpected reboot occurred, however, when the machine comes back up, it will generally restart the program again. Assuming (and yes, this is a big assumption) that the script is coded to actually check if a reboot is needed, it should continue on and complete the task. Of course you would also have to set the program to rerun on failure, or always rerun... or pretty much anything other than never rerun for this to actually occur. But from what I've seen this is the ONLY time SCCM will rerun a program without putting in multiple assignment schedules, or schedules on a timer (eg, run every day). (FYI, I found this out by accident when some of my programs were mis-configured and caused a reboot to occur when SCCM was not expecting a reboot). EDIT: The default rerun behavior for an advertisement is setup to rerun if failed, so this should be automatic. I think (but I can't be sure) that this is different from SMS 2003 behavior, which defaulted to never rerun.Anyway, I think we've beat this to death.. The core question has been answered just wanted to point out primarily a difference of opinion. And all this said, I still agree with you as to how the TS would be used in most circumstances. Scott GillSCCM Consultant

Hi everyone, thanks so much for those detailed answers. I actually LOVE the fact that there is no user interaction with task sequence and that everything has to be automated. The more I can take out of my users hands, the better (generally speaking) :)The other thing I like about task sequence is you can use "If" conditions, which to my limited knowledge, you can't do with software distribution. For instance, if I wanted to push out IE8 to my computers, but didn't want to install it onto any computers that already have IE8, I can advertise that IE8 package using task sequence.Using task sequence I am trying to execute a .js file but failing miserably. Does anyone know how I would do this?

Just clarification on deploying where software is needed / not needed. In this example, I would build a custom collection where IE8 is not already installed and then distribute it to that collection. See: http://www.myitforum.com/articles/1/view.asp?id=2559 As for the .js file, I've never tried to distribute a .js file before, but assuming it can run via a regular software distribution method, it will run with a task sequence as well (however, I do not believe WinPE has .js support... can someone else confirm that?)Scott Gill SCCM Consultant

And you would need to have that data inventoried, and have a potential delay in inventory kick of the IE deployment to machines that have IE8 already installed.Using conditional task sequences to me is a great way to implement this, and again in my personal opinion is superior to using software distribution and custom collections.Assuming you can live with the drawbacks of task sequences, and I believe all of them have been covered by now in this thread."Everyone is an expert at something" Kim Oppalfens Configmgr expert for lack of any other expertise. http://www.scug.be/blogs/sccm

blah, i was using wscript instead of cscript...... :( Thanks Jason.I tried to take the entire .js script and paste it into a single command line in TS...but the formatting is all messed up (obviously). So if I want to not push out a .js file, and instead, use the command-line task, I will need to break out each line of the .js into it's own command line, right? (told you I'm new TS) :)

blah, i was using wscript instead of cscript...... :( Thanks Jason. I tried to take the entire .js script and paste it into a single command line in TS...but the formatting is all messed up (obviously). So if I want to not push out a .js file, and instead, use the command-line task, I will need to break out each line of the .js into it's own command line, right? (told you I'm new TS) :) No, what you can do is throw the .js file into a package and use the command-line task to execute the .js file, and then point it to that package so it knows where the file is located (The option to point it to the package is right below where you enter the command line).Scott Gill SCCM Consultant

Hi Scott, Thanks again for your reply. But is there a specific reason why you would do it this way, rather than do it with a TS? Not questioning you - just trying to learn. :) Not a specific reason, but IMO it's easier to troubleshoot a single software distribution then a TS. But it looks like you already solved the problem with the .js file...Scott Gill SCCM Consultant