Zeppelin 0.7.2-1801 Release Notes

The notes below relate specifically to the HPE Ezmeral Data Fabric distribution of Apache Zeppelin. You may also be interested in the Apache Zeppelin 0.7.2 changelog and the Apache Zeppelin project homepage.

Version 0.7.2
Release Date February 2018
Source on GitHub https://github.com/mapr/zeppelin, https://github.com/mapr/livy
GitHub Release Tag 0.7.2-mapr-1801
Docker Image Name and Tags
Image Name Operating System Version of the Running Container Tag
maprtech/data-science-refinery CentOS 7.4 v1.1_6.0.0_4.1.0_centos7
Ubuntu 16.04 v1.1_6.0.0_4.1.0_ubuntu16

Zeppelin on the Data Fabric system is a component of the Data Science Refinery. This release of Zeppelin is in version 1.1 of Data Science Refinery product.

The Data Science Refinery product is packaged as a Docker container. Data Fabric ecosystem components included in the Docker image are the same as those in the EEP 4.1 release. See EEP 4.1.0 Components and OS Support for details on product version numbers.

You can run the Docker image on the following operating systems:

  • Linux (CentOS 7.x, Ubuntu 14, Ubuntu 16)
  • Windows 10 Pro (64-bit)
  • Mac OS X 10.11

The following are the verified browsers:

  • Chrome 57
  • Firefox 56.0
  • Microsoft Edge 40
  • Safari 9.0

Data Fabric product documentation is available at Zeppelin on MapR (MapR 6.1.0).

New in this Release

This release of Zeppelin on the Data Fabric system includes the following new features:

  • Support for the Spark interpreter, configured to launch Spark jobs in YARN client mode
  • Enhancements in installing custom Python environments for the Livy and Spark interpreters
  • Improvements in launching multiple Zeppelin containers on the same host

Fixes

This release by Data Fabric includes the following fixes on the base Apache release. For complete details, refer to the commit log for this project in GitHub.

The following table lists the fixes for Zeppelin:

Commit Date (YYYY-MM-DD) Comment
625acf6 2018-01-11 MZEP-94: Support for different versions of Python for the Spark interpreter
aaa8e92 2018-01-11 PACC-18: Fix regression for PACC-12
dc9cb61 2018-01-09 PACC-18: Fix incorrect home directory after docker exec run on Ubuntu
8297a9e 2017-12-25 PACC-12: Remove obsolete fix for MZEP-66
3698203 2017-12-25 MZEP-97: Enable Spark interpreter to run in YARN client mode
1e6ccc0 2017-12-14 Restore Core 6.0.0 version in Dockerfile
0418ced 2017-12-13 Update EEP and Data Fabric core versions in Dockerfiles
c6538ab 2017-12-12 MZEP-99: Fix no graphs being displayed the first time the Spark Pyspark Tutorial examples are opened
ee11e2f 2017-12-07 MZEP-98: Restore default examples in the Zeppelin Tutorial
7d15fd9 2017-12-05 MZEP-63: Explicitly set Zeppelin working directory
0fc3079 2017-12-01 Fix mapr-setup.sh URL in PACC build scripts
b8152f2 2017-11-30 MZEP-63: Enable Spark interpreter
b8e6693 2017-11-27 Refactor the script that builds Docker images

The following table lists fixes for Livy:

Commit Date (YYYY-MM-DD) Comment
bbf0155 2018-01-15 MZEP-102: Fix issue when no LIVY_RSC_PORT_RANGE is specified
90bd035 2018-01-12 MZEP-102: Provide ability to specify custom port range in livy.rsc.launcher.port.range
528e2b7 2018-01-11 MZEP-94: Support different versions of Python for the Spark interpreter
bef6778 2017-12-07 MZEP-65: Minor fix in start-in-container.sh
ccaed0f 2017-12-05 MZEP-63: Explicitly set Livy working directory
c2ddf74 2017-11-24 Fixes needed to set up a Conda environment for PySpark

Known Issues and Limitations

  • MZEP-17: The HBase interpreter cannot be used to query HPE Ezmeral Data Fabric Database Binary tables
  • MZEP-79: Legends in plots do not display correctly when running the Matplotlib (Python/PySpark) example from the Zeppelin Tutorial
  • MD-2397: Zeppelin cannot connect to Drill through the JDBC driver on a secure Data Fabric cluster when Zeppelin has Kerberos authentication enabled
  • MZEP-86: You cannot run Zeppelin as user 'root'
  • MZEP-110: You cannot use a custom R environment with Zeppelin
  • See Data Fabric PACC Known Issues for issues that apply to running Data Fabric Docker images.