AnsweredAssumed Answered

Does the hadoop fs/mfs command automatically strip path down to 'HDFS version' if originally includes NFS mounting path?

Question asked by reedv on Jul 2, 2018
Latest reply on Jul 6, 2018 by deborah

Like the question states, it seems as though the hadoop fs/mfs command automatically strips path down to 'HDFS version' if originally includes NFS mounting path. Would like to confirm this behavior, lest script something under false assumptions in the future. 

 

For example, suppose that the MapR cluster is mounted via NFS at /mapr, so MFS root can be accessed at local path /mapr/clustername/. If I then enter the command 
hadoop fs -ls /mapr/clusername/ 
it would list the root HDFS dir. just as if I had written 
hadoop fs -ls / 

 

Basically, I'd like to know if "hadoop fs [-ls | -rm | -touchz | etc.] /mapr/clustername/<some HDFS location>" will always be equivalent to  "hadoop fs [-ls | -rm | -touchz | etc.] /<some HDFS location>" .
Are there any docs that specify this behavior that I can use as a reference for how exactly this works?

What exactly is happening here? 

Outcomes