First post. Please be gentle.
In gearing up for a rollout, the admins of our enterprise (of about 5600 computers) want to drop distribution points at slow sites, most of which have only a few computers each. This seems like something SCCM is built specifically to handle, but a test run of a remote site push has revealed otherwise. We dropped a DP and five client XP workstations at the same site and subnet, and sure enough the clients ignored the local DP that had all the task sequence packages loaded up and ready to go.
I opened an MS SAS ticket, and was informed that there are only a couple of ways to ensure that a collection will pull from a specific DP:
If BITS is enabled on a DP, it takes precedence over other DP's.
Site systems can be set up as protected, and thus prevent systems outside of designated boundaries from using the protected DP.
Well, the BITS part is only helpful if we're deploying at a single site using a single DP. If BITS is enabled on two or more DP's, then we're back to computers pulling from the wrong DP.
Now, setting up protected sites seems like the perfect solution, but since our boundaries are based on sites, we don't provide fine, particulate coverage for our ravening hordes of tiny, slow remote sites. Basing our boundaries on subnets is unwiedly because we have literally hundreds of them (darn VOIP).
I'd like to keep my site boundaries and then just create subnet boundaries on demand that I could then use as a basis for creating protected sites, but this would mean that computers would fall under both a boundary for their site and a boundary for their subnet. Seems to be a consensus that having computers in overlapping boundaries is to be avoided at all costs, even when the boundaries are on the same site server.
Any ideas on how to proceed (other than go create hundreds of boundaries just to protect a relatively few sites)?
