I am trying to implement %file interpreter on zeppelin.
i have set the properties as
these properties are not working for me. Can anyone help me out with this please
How did you set up Zeppelin? There are some MapR specifics missing from the basic Zeppelin build.
we have build from zeppelin binaries
Ok... this may be a silly question...
While I have zeppelin for local development (on my Mac),
I want to set up my Mac as a client to my cluster and run Zeppelin locally but access the MapR 5.1 cluster...
Paul Curtis wrote a blog on setting up Zeppelin:
Building Apache Zeppelin for MapR using Spark under YARN | MapR
But it looks like he's building it on the cluster itself.
I'm still struggling with setting up the client. (I have some questions about that...)
However the build instructions on Zeppelin can be found:
zeppelin/README.md at master · apache/zeppelin · GitHub
So you can build for MapR.
Note: You should match the MapR cluster's version of spark (1.6.2 I think is the latest)
And you may need to specify the 2.11 version of scala.
But back to your original point...
Outside of Zeppelin, can you run the 'hadoop fs -ls ' command and see what's in your local directory?
Then it could just be making sure that Zeppelin can see your cluster.
I am not sure what the %file interpreter may allow you to do over %sh. But just wanted to bring it up as that is one of the ways I read and write to the MapR-FS, giving it works natively with NFS.
btw, there is an open discussion on a similar topic here: How to configure zepplin interpreter to access maprfs. I suggest you follow it and share your notes with the author!
Retrieving data ...