But I can 'kill -9' any stopped process that I actually stop myself (with 'kill -STOP' or <Ctrl-Z>) without 'kil -CONT', so why was this one different?
Because it's traced its asleep in the kernel. kill -9 wakes the process up and then bangs it over the head, but if it's stuck in ptrace() is can't without the debuggers say or or having been explicitly continued. I think. It's too hot.
jch
