If you have access to a root console, you can
1. determine the process ID of the hung container, eg
# ps ax | grep ctid
2. kill -9 pid
that should restore your node to normal function. if it doesnt, and the pid is zombie, you'd need to kill the node using other means- eg
echo b > /proc/sysrq-trigger
or
ipmitool power reset
if you have ipmi.