Hi, there are a couple of issues that are comingled...
My environment... A single node cluster running MapR 5.2.2 on Centos 7.4 (It works!)
My client is an iMac (also an iMac Pro) w MapR Client software installed and the copy of Zeppelin is running local to the Mac.
While I can run my scala and spark code local... I want to be able to connect to the cluster and play with MapRDB files.
Running in to a couple of issues that I am working thru, kinda hoping to be able to compare notes.
First issue, including MapR jars into Zeppelin. Its possible to set up the MapR repo and to load the jars in to a paragraph on the note, so that you can use MapR code. I have it kinda working. (I can load the specific jars if I know the full path to the jar file.) I wanted to be able to include the MapR repo by specifying the artifact information. Has anyone done this?
Note: Its possible to copy down the jars in to a local directory and to then include them in the JAR path. Being able to specify the dependencies and use the official MapR repo is a bit cooler and easier to maintain.
I'm using the %spark.dep clause and I am wondering if anyone else has played with it and MapR?
I was trying to set up SSL for connections however running in to a weird length error when Zeppelin is trying to read the ticket. But that's a different issue...