Crail file/directory permissions can be set from the config file

[raduioanstoica]

Some Spark/Hadoop applications rely on using non-default file or directory permissions, however, Crail does not implement yet this functionality. One example of a commonly used application is the Hadoop Distributed Copy tool (https://hadoop.apache.org/docs/current/hadoop-distcp/DistCp.html) that allows moving data directly between distributed filesystems. This commit allows setting custom permissions for files & directories from the Crail configuration file. It is not a proper implementation, rather a temporary fix to allow the user to change file/directory permissions and be able to run existing applications.