There is a check to see if init.sh file exists (` if [[ -e $module/init.sh ]]; then`), so it just won't get called. Regarding spark-standalone not having a init.sh that is because we dont have any initialization work to do for it (its not necessary for all modules to have a init.sh) as the spark module downloads and installs Spark.
Thanks Shivaram On Wed, Oct 8, 2014 at 2:50 PM, Nicholas Chammas <nicholas.cham...@gmail.com> wrote: > This line > <https://github.com/mesos/spark-ec2/blob/bd2b0bd89f4fe42be51b3b6431f04ca6823b43dd/setup.sh#L109> > in setup.sh initializes several modules, which are defined here > <https://github.com/apache/spark/blob/bc4418727b40c9b6ba5194ead6e2698539272280/ec2/spark_ec2.py#L575> > . > > # Install / Init module > for module in $MODULES; do > echo "Initializing $module" > if [[ -e $module/init.sh ]]; then > source $module/init.sh > fi > cd /root/spark-ec2 # guard against init.sh changing the cwd > done > > One of these modules is spark-standalone. However, it does not have an > init.sh file > <https://github.com/mesos/spark-ec2/tree/bd2b0bd89f4fe42be51b3b6431f04ca6823b43dd/spark-standalone> > . > > Should it have one? It’s the only module without an init.sh. > > Nick > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org For additional commands, e-mail: dev-h...@spark.apache.org
