I'm looking for a workaround on a problem with pxe/nfs installs (I've filed bugzilla bug 837628 on this, where there is more detail if interested). The basic problem is that the install works perfectly. However at the end, when the system should reboot, NetworkManager disconnects before the nfs-mounted install repo gets unmounted. The system then hangs forever waiting for that unmount to complete. I'm wondering whether there is anything I could do in the POST processing that would force the unmount to complete (sorry, I'm a little vague on when the POST processing happens, and on what kinds of commands are allowed there). Any suggestions greatly appreciated.

I realise this doesn't sound like a big deal - I could just manually reboot. However this problem applies to two clusters, one of which is remote from where I normally am. Having to go to that site every time I need to do a minor rebuild on one machine would be a show-stopper for F17 for me.