Spark Release 1.0.1

Spark 1.0.1 is a maintenance release with several stability fixes and a few new features in Spark’s SQL (alpha) library. This release is based on the branch-1.0 maintenance branch of Spark. We recommend users follow the head of this branch to get the most recent stable version of Spark.

You can download Spark 1.0.1 as either a source package (5 MB tgz) or a prebuilt package for Hadoop 1 / CDH3, CDH4, or Hadoop 2 / CDH5 / HDP2 (160 MB tgz). Release signatures and checksums are available at the official Apache download site.

Fixes

Spark 1.0.1 contains stability fixes in several components. Some of the more important fixes are highlighted below. You can visit the Spark issue tracker for an exhaustive list of fixes.

Spark Core

• Issue with missing keys during external aggregations (SPARK-2043)
• Issue during job failures in Mesos mode (SPARK-1749)
• Error when defining case classes in Scala shell (SPARK-1199)
• Proper support for r3.xlarge instances on AWS (SPARK-1790)

PySpark

• Issue causing crashes when large numbers of tasks finish quickly (SPARK-2282)
• Issue importing MLlib in YARN-client mode (SPARK-2172)
• Incorrect behavior when hashing None (SPARK-1468)

MLlib

• Added compatibility for numpy 1.4 (SPARK-2091)
• Concurrency issue in random sampler (SPARK-2251)
• NotSerailizable exception in ALS (SPARK-1977)

Streaming

• Key not found when slow receiver starts (SPARK-2009)
• Resource clean-up with KafkaInputDStream (SPARK-2034)
• Issue with Flume events larger than 1020 bytes (SPARK-1916)

SparkSQL Features

• Support for querying JSON datasets (SPARK-2060).
• Improved reading and writing Parquet data, including support for nested records and arrays (SPARK-1293, SPARK-2195, SPARK-1913, and SPARK-1487).
• Improved support for SQL commands (CACHE TABLE, DESCRIBE, SHOW TABLES) (SPARK-1968, SPARK-2128, and SPARK-1704).
• Support for SQL specific configuration (initially used for setting number of partitions) (SPARK-1508).
• Idempotence for DDL operations (SPARK-2191).

Known Issues

This release contains one known issue: multi-statement lines the REPL with internal references (> val x = 10; val y = x + 10) produce exceptions (SPARK-2452). This will be fixed shortly on the 1.0 branch; the fix will be included in the 1.0.2 release.

Contributors

The following developers contributed to this release:

• Aaron Davidson – bug fixes in PySpark and Spark core
• Ali Ghodsi – documentation update
• Anant – compatibility fix for spark-ec2 script
• Anatoli Fomenko – MLlib doc fix
• Andre Schumacher – nested Parquet data
• Andrew Ash – documentation
• Andrew Or – bug fixes and documentation
• Ankur Dave – bug fixes
• Arkadiusz Komarzewski – doc fix
• Baishuo – sql fix
• Chen Chao – comment fix and bug fix
• Cheng Hao – SQL features
• Cheng Lian – SQL features
• Christian Tzolov – build improvmenet
• Clément MATHIEU – doc updates
• CodingCat – doc updates and bug fix
• Colin McCabe – bug fix
• Daoyuan – SQL joins
• David Lemieux – bug fix
• Derek Ma – bug fix
• Doris Xin – bug fix
• Erik Selin – PySpark fix
• Gang Bai – bug fix
• Guoqiang Li – bug fixes
• Henry Saputra – documentation
• Jiang – doc fix
• Joy Yoj – bug fix
• Jyotiska NK – test improvement
• Kan Zhang – PySpark SQL features
• Kay Ousterhout – documentation fix
• LY Lai – bug fix
• Lars Albertsson – bug fix
• Lei Zhang – SQL fix and feature
• Mark Hamstra – bug fix
• Matei Zaharia – doc updates and bug fix
• Matthew Farrellee – bug fixes
• Michael Armbrust – sql features and fixes
• Neville Li – buf fix
• Nick Chammas – doc fix
• Ori Kremer – bug fix
• Patrick Wendell – documentation and release manager
• Prashant Sharma – bug and doc fixes
• Qiuzhuang.Lian – bug fix
• Raymond Liu – bug fix
• Ravikanth Nawada – bug fixes
• Reynold Xin – bug and doc fixes
• Sameer Agarwal – optimization
• Sandy Ryza – doc fix
• Sean Owen – bug fix
• Sebastien Rainville – bug fix
• Shixiong Zhu – code clean-up
• Szul, Piotr – bug fix
• Takuya UESHIN – bug fixes and SQL features
• Thomas Graves – bug fix
• Uri Laserson – bug fix
• Vadim Chekan – bug fix
• Varakhedi Sujeet – ec2 r3 support
• Vlad – doc fix
• Wang Lianhui – bug fix
• Wenchen Fan – optimization
• William Benton – SQL feature
• Xi Liu – SQL feature
• Xiangrui Meng – bug fix
• Ximo Guanter Gonzalbez – SQL feature
• Yadid Ayzenberg – doc fix
• Yijie Shen – buf fix
• Yin Huai – JSON support and bug fixes
• Zhen Peng – bug fix
• Zichuan Ye – ec2 fixes
• Zongheng Yang – sql fixes

Thanks to everyone who contributed!

Spark News Archive