Based on my testing, that *isn't* the case, settings in interpreter.json were not overridden by env vars.
Let me double-check and write up a recipe and get back to this. That said, though, what would *you* say would be a good way to do what I want? > On Nov 8, 2015, at 19:13, Felix Cheung <felixcheun...@hotmail.com> wrote: > > Thank you for your detailed and thoughtful mail. > > As you shall see, > https://github.com/apache/incubator-zeppelin/blob/master/spark/src/main/java/org/apache/zeppelin/spark/SparkInterpreter.java#L354 > > Env var overrides property setting, so if env var is set in zeppelin-env.sh > then the user setting the property on the Interpreter page would have no > effect. So that behavior seems to be opposite to what you want? > > _____________________________ > From: Craig Ching <craigch...@gmail.com> > Sent: Sunday, November 8, 2015 3:10 PM > Subject: Setting defaults and allowing users to override the defaults (was > Re: Setting spark.app.name?) > To: <users@zeppelin.incubator.apache.org> > > > Hey all, > > I sent my PR that allows you to configure spark.app.name, but it was > rejected. So I’m going to start a conversation on how to solve my problem in > a general way. > > I am running a service that allows users to start zeppelin instances so they > can create their own notebooks. I want to use this to expose spark (and > consequently zeppelin) to a wider audience in my organization. I have > created a separate web UI that allows users to “spin up” their own zeppelin > instance and I have some tutorials they can work through that exercise > different parts of spark so they can learn about it. I run a separate spark > cluster because I imagine I might get quite a few users and running, say, 100 > zeppelin+spark instances on a single machine just doesn’t sound like a great > idea. > > As a zeppelin site administrator, I want to be able to specify default > settings for new users, but I want to allow users to change those settings as > they become more aware of spark and zeppelin. For instance, they may want to > run their own spark cluster and run against that. > > So initially I want to be able to set spark master to my cluster and I want > to be able to specify unique spark app names so that the applications are > identifiable in the cluster. But I want users to be able to change those > settings later as they become more comfortable with spark and zeppelin. > > Just some quick comments, I don’t yet understand the need for both > zeppelin-env.sh and zeppelin-site.xml. I would actually rather have defaults > specified in zeppelin-site.xml as I’m not a huge fan of having a ton of > settings via environment variables. I don’t understand the current use of > zeppelin-site.xml, but could be used as a “first run defaults” like I need > for my use case? > > So, any thoughts? > > Cheers, > Craig > >
