I'm seeing stray knife processes on my system. I have an automated ruby script that runs bundle exec knife bootstrap against various nodes. Most of the time the knife process completes and goes away, but sometimes it stays for days. I'm noticing it days later in ps aux

I think it's related to the target node being down when knife runs. The chef server timeout is high, so the action completes eventually when the node goes back up, but I think knife may give up or hang somehow during the wait.

Is there something I can do about the stray knife processes? Does knife have timeout settings separate from the chef server's timeout settings?