Home

v4.5.5.0 Release notes

Software version

Release Date: 02/17/2020

See v4.5.5.0 for download information

Software support

  • Upgrade from 4.5.x. All that is required is an RPM upgrade.

Sensor upgrade

  • No

Certified platforms

Updates to Unravel's configuration properties

Unsupported

  • Migration Planning reports aren't supported for MapR.

New features

  • Migration Planning

    • Workload Fit

      Added this report which allows you to obtain the optimal cloud topology design (and cost estimate) corresponding to a slice of applications running on your cluster.

      Additionally it allows exercising a price-performance trade-off in determining the optimal cloud topology.

    • You can now choose Databricks as a target cloud platform for migrating to.

Improvements

  • None

Bug fixes

  • Customer:

    • Impala application specific operator counters gets ignored while trying to insert a long field to ElasticSearch. (CUSTOMER-1007)

    • Unravel now displays more information when it displays recommendations about - spark.sql.shuffle.partitions. (CUSTOMER-1169)

    • Spark application no longer hangs and fails to exit after Unravel Sensor is installed. (CUSTOMER-1201)

    • List of RBAC end user - AD groups is displayed when Business Tag groups is selected to filter the charts in Operation > Usage Detail > Infrastructure page. (CUSTOMER-1212)

    • Real user tag is shown on Applications page filter. (CUSTOMER-1213)

    • All cluster comparison metrics are compared. (CUSTOMER-1231)

    • PIG application displays duration of MR app in the Navigation tab. (CUSTOMER-1235)

    • Cluster Discovery and Forecasting reports support customized service names. (CUSTOMER-1240)

    • Cluster workload report displays consistent data. (CUSTOMER-1264)

  • Install: RPM upgrade succeed when switch to user is set to remote users. (INSTALL-357)

  • Migration Planning

    • UI displays the last generated report regardless of whether the attempt failed. The time report was generated is displayed. (REPORT-667)

    • Cluster Discovery

      • You can pick the date and time, for the report.

        The data in the charts displayed By App Type, By User and By Queue on the Cluster Discovery Report match the data in the Chargeback reports. (REPORT-684)

      • Removed mismatch in the selected Date and Time ranges between the Applications page and the Cluster Discovery report when the chosen interval is the last 7, 30, or 90 days. (UIX-2085)

    • Services and Version Compatibility:

      • Supported on CDH versions 6.1.1 and 6.0.1. (REPORT-889)

      • The latest target versions for EMR, HDI, and Dataproc are included. (REPORT-1094)

    • Cloud Mapping per Host

      • Lift and Shift: DFS Size Recommendation is consistent between Local Storage and Object Storage options. (REPORT-1117)

      • When HDP is the source system, per host logical DFS size is captured correctly for source distribution HDP. (REPORT-990, REPORT-945)

      • Unravel no longer assumes one core on-premises is equivalent to one cloud vCPU and finds the best possible match considering performance differences. (REPORT-581)

  • Reports: File Reports and Small File reports now use Spark execution engine by default. (REPORT-1220)

  • Sessions

    • Duplicate start/stop messages removed. (SESS-281)

  • Spark: Execution tab is displayed. (USPARK-422)

Known issues

  • AutoActions

    • Badge is not displayed for MR app killed by an AutoAction. (AA-91)

    • AutoAction's Violation Badge functionality is not working for Impala queries (Running, Killed). (AA-191)

    • Unknown transformation shutdowns unravel_aa demon when incorrect policies are set in the database. (AA-266)

    • Create/Update UI bug prevents user from adding individual apps by app name (only and except modes) in long-running Impala query template. (UIX-1984)

  • Cluster optimization report is empty when one day interval is selected. (REPORT-565)

  • Data insights page

    • Hive table details page parameters are inconsistently null. (Datapage-189, UIX-1874)

    • Created and Accessed Partition details are missing in Overview. (DATAPAGE-109)

    • After an upgrade from 4.5.0.x to 4.5.1.x the Table KPI (total size of all tables) data is missing. (DATAPAGE-133)

      Workaround

      1. Access the database command line: /usr/local/unravel/install_bin/db_access.sh

      2. Truncate dashboard summaries table: mysql> delete from dashboard_summaries where entity_type='dkt';"

      3. Restart Unravel tw daemon: /etc/init.d/unravel_tw restart

    • When switching view by labels (All, Warm, Cold, and Hot) the page can become unresponsive. (UIX-1878)

      Workaround

      • Only solution available is to relaunch the crashed browser, and toggle ONLY ONCE, from default view ALL to any of the following labels (Hot, Warm, Cold).

  • HBase: Namespace missing in HBase table name. (UIX-1736)

  • Kafka isn't working when clusterID has a space. (UKAFKA-69)

  • Migration planning

    • Migration planning reports are not supported for MapR

    • The following Cloud product/service configurations are not supported:

      • Cloud Product/Service: Azure (IaaS), Azure (HDI)

        Storage Type: Object

        Region: Germany Northeast, South Africa North, Germany Central, US DoD East, US DoD Central, South Africa West, US Gov Virginia, US Gov Iowa.

    • Cluster Discovery Report

      • OS value is not shown for source distribution CDH 6.3. (REPORT-930)

    • Services and Version Compatibility:

      • Some of the latest versions of the components shown may be different from those present in the actual source environment. This is because the ones shown are based on the Distribution Stack's Version and not captured from the actual environment itself.

    • Cloud Mapping per Host

      • The New Report modal hangs when you don't provide the Unravel property for cluster type (CDH/HDP) com.unraveldata.cluster.type. (REPORT-731)

      • Lift and Shift: DFS Size Recommendation is different between Local Storage and Object Storage options. (REPORT-1117)

      • Unravel assumes one core on-premises is equivalent to one cloud vCPU and finds the best possible match without considering any performance differences. (REPORT-581)

      • Compatibility between selected VM type and the storage tier/class isn't verified at this time. (REPORT-940)

      • Scheduled report fails to generate a report.

        Workaround:

        Generate a report as and when needed on-demand. (REPORT-1140)

    • Cloud Mapping per Host, Workload Fit

      • Resetting/updating custom cost values doesn't work correctly. (UIX-2135, REPORT-1316, REPORT-1312).

    • Workload Fit

      • Single Cluster with auto-scaling doesn't take into consideration the minimum number of master node(s) needed. (REPORT-1018)

  • MR:

    • The Move to Queue option in Actions (MR APM) does not work. (PLATFORM-1195)

    • Cluster ID missing for MR apps. (PLATFORM-1434)

    • Timestamp is missing for Failed MR apps. (UIX-1873)

    • Not able to select last block of Histogram graph in timeline page(/UIX-2388)

  • Platform

    • appstatus.AppInfoAccessor: Unable to update ES document. (CUSTOMER-658)

    • Spark worker is missing Azure libraries which causes failure when loading event log and executor logs from Azure.

    • RM Polling is delayed when Yarn kill and move actions are enabled in AutoActions. (PLATFORM-1479)

    • Setting impalad as Impala data source raises errors for invalid HTTP address. (PLATFORM-1444)

  • Queue Analysis: Graph zooming and resetting is not working properly in Edge. (UIX-158)

  • Small Files etl_fsimage process fails while trying to insert into the Hive Table. Workaround: use Spark to process the FSImage.

  • Sessions

    • Limitation in HDP cluster: While applying recommendations on a Spark session sometimes we get error like “Could not run the Spark application on the cluster: java.lang.IllegalStateException: hdp.version is not set while running Spark under HDP, please set through HDP_VERSION in spark-env.sh or add a java-opts file in conf with -Dhdp.version=xxx”. The exact reason for this bug is yet to be ascertained.

      Workaround

      1. Open and edit spark-env.sh of the client where unravel is running.

      2. Add a line similar to export “HDP_VERSION=2.6.5.0-292". Make sure to put the proper HDP version in the line.

    • In a MapR cluster, Spark apps fail after tuning recommendations are applied. (SESS-250)

    • Hive: Newly tuned app in MapR is not shown after applying recommendations. (SESS-251)

  • Spark

    • Query plan and query text do not match in the Spark APM. (DT-186)

    • Some of the applications doesn't display both Query and Query Plan. (USPARK-511)

    • User is not directed by back to the Unravel UI when cancelling print chart. (UIX-2454)

  • UIX

    • Global search and filter by app name are not working. (UIX-1998)

    • Download CSV and XLS is not working for Tez and Spark APMs. (UIX-1999)

  • Workflow: name search is case-sensitive. (UIX-1904)

For support issues contact Unravel Support.

In this section: