Enterprise Steam Release Notes¶
Version 1.6.5 (Mar 2, 2020)¶
- Fixed custom Hadoop temporary directory to be part of HADOOP_CLIENT_OPTS
Version 1.6.4 (Feb 12, 2020)¶
- Fixed error in Hadoop Helper for Java 7 Hadoop distributions
- Fixed error when creating Python environment with the same name as deleted environment
- Improved validation and documentation for LDAPs with internal certificate authority
- Improved documentation for Driverless AI config.toml override
- Improved documentation for Driverless AI installation prerequisites
Version 1.6.3 (Jan 31, 2020)¶
- Supports RSparkling 3.28
- Use Hadoop Tmp Dir for Hadoop Helper commands to prevent slowdowns
- Fixed an issue with old DAI servers preventing instances from loading
- Fixed server panic with invalid LDAP over TLS configuration
- Fixed initial LDAP config requiring server restart
Version 1.6.2 (Jan 20, 2020)¶
- Supports Sparkling Water 3.28
- Fixed LDAP over TLS
- Fixed SAML token login in R client
- Fixed H2O profile not saving YARN queues
Version 1.6.1 (Jan 14, 2020)¶
- Fixed mutual TLS setup
- Fixed Driverless AI version parser
- Add datatable_strategy WRITE to default DAI launch config
- Add Driverless AI instance password to the instance details
- Docs: Updated copyright year to include 2020
Version 1.6.0 (Dec 18, 2019)¶
- Complete UI overhaul of configuration page for administrators.
- New administrator settings for individual products and deployment backends and other configuration.
- Improved UI/UX experience.
- Moved most of the settings from steam.yaml into the administrator configuration in Steam UI.
- Configuration file steam.yaml is no longer required.
- Enabling Hadoop backend now thoroughly validates all settings to ensure they are correct (Proxyuser, Kerberos, …).
- Added ability to export and import configuration.
- Several settings no longer require Enterprise Steam restart.
- Added indicator when Enterprise Steam needs to restart to apply changes.
- Added option to use existing Livy server for Sparkling Water as long as it is on the same host as Enterprise Steam.
- HAProxy comes bundled and does not need to be installed separately.
- Steam Hadoop Helper (Steam Manager) comes bundled and does not need to be installed separately.
- Additional HADOOP_CLASSPATH options can be added when launching H2O clusters.
- Added ability to download complete Enterprise Steam logs from the UI.
- Added SAML configuration validation.
- Added cacert option when connecting to Steam or H2O clusters via Python client.
- Added option to ensure internal secure connections with H2O and Sparkling Water clusters.
- Added warning when uploading non-Hadoop version of H2O engine.
- Improved database migration.
- Improved validation of H2O Launch Parameters.
- Fixed bug where Sparkling Water could not launch when users reused cluster names.
- Fixed bug where Sparkling Water could not launch when the cluster name contained spaces.
- Fixed UI issue when uploading Python environments with the same name.
- Fixed Python environment selection when no Python environments assigned.
- Fixed auto-assigment of profiles to LDAP groups
Version 1.5.1 (Nov 21, 2019)¶
- Fixed a database migration bug affecting new installations.
Version 1.5.0 (Oct 10, 2019)¶
- Added ability to launch Driverless AI instances.
- Added ability to use Conda environments with Sparkling Water
Version 1.4.11 (Oct 1, 2019)¶
- Fix token authentication in Python client when SAML is used
- Fix R and Python client when internal and external ports do not match
- Support recent versions of Sparkling Water
Version 1.4.10 (Jul 19, 2019)¶
- Added an ability to download H2O logs from Steam UI
- Added a config option for custom KRB5_CONFIG (STEAM_KRB5_CONFIG)
- Added support for templating in YARN queue names
- Added a check to ensure the Spark version and Sparkling Water version match
- Added a check to warn admin if Steam port is occupied
- Improved fetching of H2O YARN job status to now use impersonation
- Fixed h2osteam Conda package
- Fixed issue when Sparkling Water cluster failed to start when SPARK_YARN_MODE property was set
Version 1.4.9 (May 21, 2019)¶
- Added Conda packages for Python client
- Added ability to delete multiple stopped clusters at the same time
- Added Python PySpark path configuration to profile
- Added YARN queue configuration to user setting
- Removed the need for secondary external port
- Configuration fields are now protected from trailing and leading whitespace characters
- Updated pysparkling dependencies for Python2
- Fixed an issue with YARN usernames being lowercased when they were not configured to do so
Version 1.4.8 (Apr 15, 2019)¶
- Add support for RSparkling through Jupyter notebook
- Add option to disable Jupyterhub
- Expose internal address and node ID for H2O cluster in the Steam Web UI
- Add option to specify path to custom Pyspark Python for Sparkling Water
- Fix pysparkling dependencies for Python3
- Updated engine upload backend
Version 1.4.7 (Apr 3, 2019)¶
- Fix Sparkling Water proxy issue with upercase usernames
- Improve uploading h2o-3 engines
- Set SPARK_YARN_MODE correctly based on the hadoop distribution
Version 1.4.6 (Apr 1, 2019)¶
- Added ability to choose H2O-3 Leader Node when starting a cluster
- Added ability to control the number of clusters a user can spin per cluster profile
- Added option to select default Sparkling Water backend
- Added automatic redirection back to login with an expired session cookie
- Added an ability to auto-assign Steam profiles according to SAML profiles
- Docs: Add “Before you begin installation” section
- Docs: Documented steam.yaml configuration options
- Docs: Updated documentation
- Fix an issue when Steam was hitting API endpoints of dead clusters
- Fix and issue when hadoop-unjar files were not deleted from temp directory
- Fix issue with upercase usernames and Sparkling Water on Hadoop
Version 1.4.5 (Mar 22, 2019)¶
- Added Configurable Steam Web UI timeout (STEAM_WEB_UI_TIMEOUT_MIN)
Version 1.4.4 (Feb 20, 2019)¶
- Make log file permissions configurable (STEAM_LOG_PERMISSIONS)
- H2O: Communicate with cluster using leader node only
- SW: Added support for Hive tables
- SW: Disable Spark dynamic allocation for internal backend
- SW: Bundle and distribute all pysparkling dependencies
- LDAP group configuration is no longer mandatory
- Bug fixes for Jupyterhub
- Bug fixes for Sparkling Water params
- Bug fixes for CDH5
Version 1.4.3 (Jan 29, 2019)¶
- Added ability to provide secure HTTP headers (SERVER_STRICT_TRANSPORT, SERVER_X_XSS_PROTECTION, SERVER_CONTENT_SECURITY_POLICY)
- Improved logging output, time in UTC
- Accept encrypted TLS private keys
- Log files permissions elevated for the group
- Bug fixes for HDP Spark
- Bug fixes for Sparkling Water integration
- Bug fixes for Steam Python API
- Bug fixes for LDAP connector
- Bug fixes for Jupyterhub
- Bug fixes for Steam Web UI
Version 1.4.2 (Jan 1, 2019)¶
- Add ability to upload H2O engine jar from Python/R API
- Remove cluster dialog waits until cluster has been shut down
- Log successful logins
- Log files permissions reduced
- Steam nbextension shows “please wait” when changing clusters
- Document setup and troubleshooting steps for Hadoop/Kerberos configuration
- Bug fixes
Version 1.4.1 (Dec 3, 2018)¶
- Enable Sparkling Water API from API
- Hide disabled user with a checkbox
- Display message if there are no clusters to show
- Fix glibc dependency for steam binary
- Fix documentation version
- Add release notes to documentation
Version 1.4.0 (Nov 23, 2018)¶
- Add Sparkling Water integration
- Add SAML authentication
- Add More detailed cluster profiles
- New cluster overview
- New launch cluster page
- Add option to generate Personal access tokens