I refreshed my Spark version to the master branch as of this morning, and am noticing some strange behavior with executors and the UI reading executor logs while running a job in what used to be standalone mode (is still now called coarse grained scheduler mode or still standalone mode?). For starters, the job seems to spin off only 1 executor (# 0) as per the job's console log
However, when I drill down to the job's executor logs in SPARK_HOME/work/<job-id>/, I see two directories "0" and "1". I was expecting to see only the "0" directory since the job registered only a single executor. I notice that my job's log output are sent to the stdout in the "0" directory. Secondly, the stdout/stderr links on the web UI seem to point to executorId=1, and since my job's output is sent to the 0/stdout, the logs on the UI don't have my job's log output. Is this a known issue, or am I missing something? Thanks, Ameet
