I'm trying to run a bash script located at a HDFS folder via oozie workflow. This script has a lot o configurations to be done by the user, and they are passed to the script via a file the script has to read and is put at the same HDFS folder as the script. The script cannot read it, though, and I wonder if I should comply with some specific path configurations for this to work on HDFS.
When this code here is reached:
if [ ! -e `pwd`/vars.sh ] ; then
echo "Unable to continue: file vars.sh doesn't exist!"
the script exits. Is the "pwd" a problem? I know that it pressuposes that the oozie workflow would respect the path of execution of my script. Is this a mistake?