Hey All, Today we merged a pull request which changes the organization of the scripts use to run Spark. The change is fairly minor but it will affect anyone who packages spark or uses spark's binary scripts. Spark's scripts are now divided into a /bin and /sbin directories. This will be the layout in Spark 0.9.
/bin contains user-facing scripts: /bin/spark-shell /bin/pyspark /bin/spark-class /bin/run-example /sbin contains administrative scripts for launching the standalone cluster manager: /bin/start-master.sh /bin/start-all.sh ...etc This allows administrators to set permissions differently for user and admin scripts. It is also consistent with the way many other projects package binaries. This change was contributed by @xiajunluan, @shane-huang, and @ScrapCodes. https://github.com/apache/incubator-spark/pull/317 Cheers, - Patrick
