I have the same issue on my production system. While I used the 2.6.24 kernel with PVE 1.6 everything worked fine. Since I have upgraded system to the latest 2.6.32 PVE kernel and ProxMox 1.7, some containers begin to crash in the similiar way. If the problematic container "hangs up", the commands "vzps" and "vztop" also crash with segmentation fault errors. In this cases I am forced to reboot the whole host-machine. Furthermore, I have to reboot it in incorrect way, by turning down its power, because the vz processes refuse to stop gracefully. The command "vzstop" does not have success as well and the process of "vzstop" remains running even the command "kill -9" is issued against it. "vzctl chkpnt $VEID --kill" cant help me too.
I observe this strange behaviour on two different boxes and on random containers. So, this is not hardware troubles. I repeat, before the kernel upgrade from 2.6.24 to 2.6.32 everything was fine. I have met no serious problems with ProxMox/OpenVZ for the two past years. And now, I don't know what to do. Because 2.6.24 is not supported anymore, and 2.6.32 is enough raw and buggy yet.