Skip to main content


v4.4.0 Release notes

Software version

Release Date: 09/17/2018 for 4.4

For details on downloading updates see here. Hotfix

Certified platforms

Please also review the Compatibility Matrix before upgrade/install.

  • CDH: On-premise CDH 5.12, 5.13 & 5.14 (incl. Spark 2.3.x) with Kerberos.

  • HDP: On-premise HDP with Kerberos & spnego enabled.

  • MapR: 6.0.0 with MapR Expansion Packs 4.1.1.

Unravel sensor upgrade

  • Not Applicable for Fresh Installation

New features

  • Reports

    Multiple reports have been added to enable to you manage and tune your cluster and its resources. You must have OnDemand installed and use MySQL to use these features.

    • You can schedule reports to run at specific times and days.

    • All reports are archived and available to be viewed or downloaded until you delete them.

    • Cluster Optimization - generates cluster wide default recommendations to tune your cluster. Hive and MapReduce is currently supported.

    • Forecasting – uses historical disk capacity usage to extrapolate future disk needs.

    • Small Files – analyzes your directories, using your file size criteria, to help manage costly small file proliferation.

  • Impala

    • Insights have been greatly enhanced.

    • Auto actions templates for long-running Impala queries and rogue Impala HDFS RW have been added.

  • HBase Cluster - Support for HBase Cluster has been added.

  • Sessions - Ability to tune/auto tune your applications.

  • PostgreSQL 10.2.1 is now the default database embedded in the RPM

Improvements and bug fixes

  • Improved detection of unresponsive YARN applications. (PLATFORM-659)

Bug fixes
  • Customer fixes

    • Failed jobs are only shown if Application ID is available (CUSTOMER-192)

    • Cluster Workload Report populates correctly. (CUSTOMER-209, CUSTOMER-211)

    • Reduced unravel_ngui daemon startup time. (CUSTOMER-244)

    • Enabled "search" by app name field on Applications page tabs. (CUSTOMER-263)

    • Removed unnecessary fields from MapReduce Timeline view. (CUSTOMER-279)

    • Tez tagged workflows are now working correctly. (CUSTOMER-272, CUSTOMER-282)

  • Kafka

    • Added the ability to show more than one active broker at a time (CUSTOMER-93)

    • Need the ability to show Kafka graphs aggregated at cluster level not per broker (UIX-660)

  • is disabled for all unravel processes due to JMX vulnerability concerns. (PLATFORM-382)

  • Spark

    • Spark UI Navigational issues are fixed (USPARK-122,USPARK-127)

    • Spark Execution DAG arrow direction is fixed (USPARK-120)

    • Spark Event will be shown only if Events are available (USPARK-124)

  • Tez

    • Various improvements in Tez pipeline efficiency (TEZLLAP-110)

Known issues

  • YARN applications might be reported as RUNNING even though they are still waiting to be scheduled. (PLATFORM-715)

  • Global search returns multiple results for single Hive Query ID (UIX-1224)

  • CMP API issue picks up data ~4 days later than the earliest data point viewable (SESS-60)

  • AutoActions are not triggering for long running Hive apps with Tez as Execution Engine (PLATFORM-643)

Configuration properties

See here.

Software Upgrade Support

  • New install only; upgrades are not supported. Hotfix

Release Date: 10/11/2018

This hotfix addresses the below issues.

Issue ID





Tez applications don't show up in Unravel UI.



Event Worker will crash when "com.unraveldata.yarn.timeline-service.webapp.address" is not configured.



LDAP Authentication fails when hive.server2.authentication.ldap.groupFilter is configured.


Hive Sensors

Hive apps don't show up due to differences in API return values in FileSinkDesc API

For support issues, contact Unravel Support.