[discuss] if runners permission is empty, every authenticated user can view the notebook

Hi Folks, https://issues.apache.org/jira/browse/ZEPPELIN-4167 if runners permission is empty, every authenticated user can view the notebook. I also think that if you don't set permissions, you shouldn't use anyone. This is more rigorous, Mor

Re: [discuss] if runners permission is empty, every authenticated user can view the notebook

I think the default value should be the user who created the note. Xun Liu 于2019年5月23日周四 下午10:28写道： > Hi Folks, > > https://issues.apache.org/jira/browse/ZEPPELIN-4167 > if runners permission is empty, every authenticated user can view the > notebook. > > I also think that if you don't set perm

Re: [discuss] if runners permission is empty, every authenticated user can view the notebook

If I’m understanding correctly, this is exactly how we want zeppelin to behave in our environment: We want all notebooks to be viewable by anyone who can authenticate into our zeppelin server. I would hate to lose this feature or suddenly be required to change permissions on every created notebo

Re: [discuss] if runners permission is empty, every authenticated user can view the notebook

Do it via configuration make sense to me. Paul Brenner 于2019年5月23日周四 下午10:54写道： > If I’m understanding correctly, this is exactly how we want zeppelin to > behave in our environment: We want all notebooks to be viewable by anyone > who can authenticate into our zeppelin server. I would hate to

Spark job fails when zeppelin.spark.useNew is true

Hello - I am looking for insight into an issue I have been having with our Zeppelin cluster for a while. We are adding a Geomesa-Accumulo-Spark jar to the Spark interpreter. The notebook paragraphs run fine until we try to access the data, at which point we get an "Unread Block Data" error from the

Re: Spark job fails when zeppelin.spark.useNew is true

We use Geomesa on Accumulo with Spark and Zeppelin on a Kerberized cluster (hdp3). We've had a number of issues, but that one doesn't look familiar. >From memory, we had to: Build geomesa spark with Accumulo version to match our cluster, and libthrift to match Accumulo, and another version change

Re: Spark job fails when zeppelin.spark.useNew is true

>>> adding a Geomesa-Accumulo-Spark jar to the Spark interpreter. How do you add jar to spark interpreter ? It is encouraged to add jar via spark.jars Krentz 于2019年5月24日周五 上午4:53写道： > Hello - I am looking for insight into an issue I have been having with our > Zeppelin cluster for a while. We

Re: Spark job fails when zeppelin.spark.useNew is true

I add the jar by editing the Spark interpreter on the interpreters page and adding the path to the jar at the bottom. I am not familiar with the spark.jars method. Is there a guide for that somewhere? Could that cause the difference between spark.useNew being set to true versus false? On Thu, May

Re: Spark job fails when zeppelin.spark.useNew is true

spark.jars and spark.jars.packages is the standard way to adding third party libraries. And it works for all the native supported modes (standalone/yarn/mesos and etc). The approach you used only works for old spark interpreter, and is not a standard way to adding jars for spark engine (e.g. it won

Re: [discuss] if runners permission is empty, every authenticated user can view the notebook

Hi Paul, The default value of Reader is all the users make sense for me, but regarding the Writer/Runner, it seems better to be the owner user by default. Will that affect your existing installation ? Paul Brenner 于2019年5月23日周四 下午10:54写道： > If I’m understanding correctly, this is exactly how we