Hi Folks, a heads up, to stop you making the same mistake as me.
I have just spent the last three weeks finishing up an OpsMgr rollout for a large manufacturing company in the UK as part of a server re-location project, and as they have a pretty well deployed SCCM infrastructure, I opted for AD Integrated for Agent deployment with the agent being delivered by SCCM.
The target for the agent was all Windows Servers in the domain, which should have made the agent roll-out pretty simple, however, I forgot to instruct the SCCM engineer to make an exception for my RMS, two Management Servers and two ACS Collectors, and he managed to physically deploy and agent to each of these boxes, the result was pretty catastrophic, as the Health Service on the MS’s simply stopped working. The action you may think was most obvious here was to roll back the agent but this actually resulted in ripping the guts out of the MS’s with all services refusing to start (even after a repair).
Luckily, I managed to pretty much re-install all of the MS’s back to their previous states but obviously this has meant re-configuring things like ACS and has resulted in a pretty significant outage for the client.
Moral of the story, make sure you make an exception for your OM infrastructure servers if planning an agent deployment via SCCM
Note for Microsoft, please make the OpsMgr agent or SCCM intelligent enough to know that a server is running a Management Server component..........
Surely this bit of logic should sit in the MSI ... if i'm already installed, and it's an MS then don't install? Shame on those OpMgr code monkies!
Sounds to me that ConfigMgr did the job spot on! Hoorah!