Past Releases


New features

  • New Slice API:
    • Add the ability to create IndexPartition based on the desired number of documents per split #812 (issue: #778)
  • Spark:
    • Support for Spark 2.0 / Datasets #647

Enhancements

  • Rest:
    • Validate field exclusion settings when EsSpark.saveJsonToEs() is executed #782
    • Review the error messages in ES 5.x and properly parse them #779

Bug fixes

  • Rest:
    • Fixed parsing of ES mapping with OBJECT field named 'properties' #810 (issue: #809)
    • Not all errors in ES are reported when writing #720
  • MR and Spark:
    • Restore data locality preference for hadoop and spark #819 (issue: #814)
  • Spark:
    • saveToEs saves fields with NULL values #792
  • Yarn:
    • Authenticate with a local keytab in the YARN Application Master #807

Non-issue

  • Spark:
    • Got exception when I tried to invoke _mtermvector #796
    • ES-Hadoop on Spark 2.0 #759

Documentation

  • HDFS Repository:
    • Repository HDFS doc improvements #781
    • Snapshot Creation Exception #813
  • Rest:
    • Corrected property name #805

Known Issues

  • Core ES-Hadoop:
    • Restore data locality preference for hadoop and spark #814
    • Update/Upsert bulk actions with scripts are broken for ES 1.x as of 5.0.0-alpha5 #817
Click Here

Detailed information on this release is available in the release notes.

Click Here







Enhancements

  • Validate field exclusion settings when EsSpark.saveJsonToEs() is executed #782

Bug fixes

  • Fixed parsing of ES mapping with OBJECT field named 'properties' #810 (issue: #809)
  • Authenticate with a local keytab in the YARN Application Master #807

Documentation

  • Corrected property name #805

Non Issue

  • Snapshot Creation Exception #813
Click Here




Fixed broken tile maps by switching to Elastic Tile Service and making the map provider configurable.

Click Here



Bumped bundled node.js version to 4.4.7

Click Here

Enhancements

  • Rest:
    • Align extended Boolean parsing with Elasticsearch #798
    • Replace/Ignore DOT character in field names before inserting into Elastic Search #758

Bug fixes

  • MR:
    • Reading boolean 0/1 values from ES into Spark does not work (although false/true is ok) #795
  • Pig:
    • Pig maps can have values of different types #777
  • Rest:
    • es.index.auto.create setting doesn't work as expected #793
    • Parsing error messages can cause a NPE #776
  • Spark:
    • saveToEs saves fields with NULL values #792

Non-issue

  • Hive:
    • EsHadoopIllegalArgumentException: Cannot detect ES version #794
    • ES-Hive throwing exception while reading for bulk no. of columns #790
    • Create a external table with 'location' command success while elasticsearch-hadoop do not support 'location' function #786
  • Pig:
    • Store Tuple as object with Pig in Elasticsearch #746
  • Rest:
    • Elasticsearch : Cannot detect ES version #791
  • Spark:
    • elasticsearch-hadoop exception about mapping #766
    • Parameter ES_MAPPING_TIMESTAMP is not working using saveToEs #765

Documentation

  • General:
    • Provide Kerberos configuration documentation #762
    • Update cloud.adoc #773
    • Update configuration.adoc #772
  • Hive:
    • corrected link to 'ADD' command #788
  • Spark:
    • Fix example for spark. #768 (issue: #767)
Click Here