If you issue a "mapred job -kill" against a uberized job, the job (and the yarn application) state transitions to KILLED, but the application master process continues to run. The job actually runs to completion despite the killed status.
This can be easily reproduced by running a sleep job:
Issue a kill with "mapred job -kill [job-id]". The UI will show the job (app) is in the KILLED state. However, you can see the application master is still running.