This is the documentation for CDH 4.7.0.
Documentation for other versions is available at Cloudera Documentation.

What's New in CDH4.3.0

Apache HDFS

  • New capability for balancing usage across a DataNode's disks. (HDFS-1804)
  • The default client maximum heap size (set in HADOOP_CLIENT_OPTS) has been raised from 128 MB to 512 MB.

Apache Avro

  • New version: Avro 1.7.4

Apache Flume

  • RPC using Thrift (FLUME-1894).
  • Communication between AvroSource and NettyAvroRpcClient can be compressed by means of ZlibEncoder and ZlibDecoder (FLUME-1915).
  • Flume can write two checkpoints so that the logs do not need to be replayed in case of a failure or shutdown while a checkpoint is being written (FLUME-1516).

Apache HBase

  • New version: HBase 0.94.6.1
  • Fixed a problem that occurred when you attempted to split a region with no store files. Now when you apply a split to an empty region, it is split into two empty regions (HBASE-7876).

Apache HCatalog

  • New version: HCatalog 0.5.0

Hue

  • New metastore editor application (MetaStore Manager) provides Hive metastore information and operations (list/drop tables, partitions...)
  • Support for HDFS trash functionality (HUE-995).
  • HiveServer 2 integration allows you to use either Beeswax or Hive Server 2 as a backend.
  • Support for Hive and Impala syntax highlighting and autocompletion
  • Support for Oracle 11.2
  • Support for multiple statements per query (HUE-995).
  • New Pig editor application.
  • Support for Oozie bundles.
  • Support for Beeswax and Impala queries, Job Designer desgins, Oozie workflows, coordinators, bundles trash

Apache Oozie

  • New version: Oozie 3.3.2
  • suspend and resume options allow you step through Oozie workflow actions (OOZIE-1245) .

Apache Pig

  • New version: Pig 0.11.0

Apache Sentry (incubating)

Sentry enables role-based, fine-grained authorization for HiveServer2 and provides classic database-style authorization for Hive and Cloudera Impala. Sentry is provided as part of CDH4.4, but a standalone version is available as of CDH4.3.0.

Installing the Standalone Version of Sentry (Provided with CDH4.3.0)

  Important: Do not do this for CDH4.4 or later —Sentry is included with CDH4.4. See What's New in CDH4.4.0.

To download and install the standalone version of Sentry provided with CDH4.3.0, follow instructions under If you want to install the version of Sentry provided with CDH4.3.0.

To configure the standalone version of Sentry manually, follow the instructions under Configuring Sentry; the configuration steps are the same whether you are using CDH4.3 or CDH4.4.

To configure the standalone version of Sentry using Cloudera Manager 4.5 or 4.6, see If you are using Cloudera Manager 4.5 or 4.6.

Apache Sqoop

  • New version: 1.4.3
  • Sqoop no longer serializes database-connection credentials into MapReduce job-configuration objects, so this information is no longer visible to everyone.

Apache Sqoop 2

  • Sqoop 2 no longer serializes database-connection credentials into MapReduce job-configuration objects, so this information is no longer visible to everyone.