Rudder

This is due to a race condition in the lock handling:
- cfengine interprocess locks are stored in lmdb
- when cf-serverd starts, it holds a lock and releases it on stop
- when it restarts, it checks this lock and if it expired, it tries to kill the owning process
- sometime with pid turnover, it can be possible that this pid has been reused by another process in the meantime

We should add a check to make sure we only ever kill cfengine processes