I finally had a chance to take a look at the frustrating machine from a couple of posts ago - pulled the video card out and put it back in again, now it's fine. (The hard disk was fine, although a previous disk from that machine that had apparently failed actually was dead.)
We also took delivery of most of the clusters that I'll be "getting" to set up - 50ish compute nodes and 3 head nodes. No racks or switches for them yet though. Pictures here - we're temporarily housing them in the machine room belonging to an entirely different research group. No technicaly details yet - we don't really know what we're running on them. This is the stuff I alluded to in an earlier post.