After adaqfs came back, Alex and I rebooted the counting house machines (mostly to get rid of stale NFS handles) and temporarily reconfigured them to use the central name servers only. (DNS on adaqfs is not yet restored.) All this went smoothly. Everything seems to be back up and running right now, at least from the systems point of view. ROOT and Podd on /adaqfs/apps is up-to-date (Podd version 1.5.16-dvcs from Oct 29). So far so good.
What's left to do is to check DAQ and various automation scripts - whatever was run from the home directories on adaqfs. (System scripts shouldn't be on a remote server, but who knows.)