Release Notes - Hadoop Common - Version 0.10.0 - HTML format

Bug

  • [HADOOP-546] - Task tracker doesnt generate job.xml in jobcache for some tasks ( possibly for only rescheduled tasks)
  • [HADOOP-596] - TaskTracker taskstatus's phase doesnt get updated on phase transition causing wrong values displayed in WI
  • [HADOOP-628] - hadoop hdfs -cat replaces some characters with question marks.
  • [HADOOP-629] - none of the rpc servers check the protcol name for validity
  • [HADOOP-673] - the task execution environment should have a current working directory that is task specific
  • [HADOOP-700] - bin/hadoop includes in classpath all jar files in HADOOP_HOME
  • [HADOOP-737] - TaskTracker's job cleanup loop should check for finished job before deleting local directories
  • [HADOOP-738] - dfs get or copyToLocal should not copy crc file
  • [HADOOP-744] - The site docs are not included in the release tar file
  • [HADOOP-745] - NameNode throws FileNotFoundException: Parent path does not exist on startup
  • [HADOOP-752] - Possible locking issues in HDFS Namenode
  • [HADOOP-764] - The memory consumption of processReport() in the namenode can be reduced
  • [HADOOP-770] - When JobTracker gets restarted, Job Tracker History doesn't show the jobs that were running. (incomplete jobs)
  • [HADOOP-774] - Datanodes fails to heartbeat when a directory with a large number of blocks is deleted
  • [HADOOP-777] - the tasktracker hostname is not fully qualified
  • [HADOOP-782] - TaskTracker.java:killOverflowingTasks & TaskTracker.java:markUnresponsiveTasks only put the tip in tasksToCleanup queue, they don't update the runningJobs
  • [HADOOP-786] - PhasedFileSystem should use debug level log for ignored exception.
  • [HADOOP-792] - Invalid dfs -mv can trash your entire dfs
  • [HADOOP-794] - JobTracker crashes with ArithmeticException
  • [HADOOP-802] - mapred.speculative.execution description in hadoop-defauls.xml is not complete
  • [HADOOP-813] - map tasks lost during sort
  • [HADOOP-814] - Increase dfs scalability by optimizing locking on namenode.
  • [HADOOP-818] - ant clean test-contrib doesn't work
  • [HADOOP-823] - DataNode will not start up if any directories from dfs.data.dir are missing
  • [HADOOP-824] - DFSShell should become FSShell
  • [HADOOP-825] - If the default file system is set using the new uri syntax, the namenode will not start
  • [HADOOP-829] - Separate the datanode contents that is written to the fsimage vs the contents used in over-the-wire communication
  • [HADOOP-835] - conf not set for the default Codec when initializing a Reader for a record-compressed sequence file
  • [HADOOP-836] - unit tests fail on windows (/C:/cygwin/... is invalid)
  • [HADOOP-838] - TaskRunner.run() doesn't pass along the 'java.library.path' to the child (task) jvm
  • [HADOOP-840] - the task tracker is getting blocked by long deletes of local files
  • [HADOOP-841] - native hadoop libraries don't build properly with 64-bit OS and a 32-bit jvm
  • [HADOOP-844] - Metrics messages are sent on a fixed-delay schedule instead of a fixed-rate schedule
  • [HADOOP-846] - Progress report is not sent during the intermediate sorts in the map phase
  • [HADOOP-849] - randomwriter fails with 'java.lang.OutOfMemoryError: Java heap space' in the 'reduce' task

New Feature

  • [HADOOP-454] - hadoop du optionally behave like unix's du -s
  • [HADOOP-574] - want FileSystem implementation for Amazon S3
  • [HADOOP-681] - Adminstrative hook to pull live nodes out of a HDFS cluster
  • [HADOOP-811] - Patch to support multi-threaded MapRunnable

Improvement

  • [HADOOP-331] - map outputs should be written to a single output file with an index
  • [HADOOP-371] - ant tar should package contrib jars
  • [HADOOP-451] - Add a Split interface
  • [HADOOP-470] - Some improvements in the DFS content browsing UI
  • [HADOOP-524] - Contrib documentation does not appear in Javadoc
  • [HADOOP-525] - Need raw comparators for hadoop record types
  • [HADOOP-571] - Path should use URI syntax
  • [HADOOP-618] - JobProfile and JobSubmissionProtocol should be public
  • [HADOOP-619] - Unify Map-Reduce and Streaming to take the same globbed input specification
  • [HADOOP-621] - When a dfs -cat command is killed by the user, the correspondig hadoop process does not get aborted
  • [HADOOP-676] - JobClient should print user friendly messages for standard errors
  • [HADOOP-717] - When there are few reducers, sorting should be done by mappers
  • [HADOOP-720] - Write a white paper on Hadoop File System Architecture, Design and Features
  • [HADOOP-756] - new dfsadmin command to wait until safe mode is exited
  • [HADOOP-763] - NameNode benchmark using mapred is insufficient
  • [HADOOP-783] - Hadoop dfs -put and -get accept '-' to indicate stdin/stdout
  • [HADOOP-796] - Node failing tasks and failed tasks should be more easily accessible through jobtracker history.
  • [HADOOP-804] - Cut down on the "mumbling" in the Task process' stdout/stderr
  • [HADOOP-806] - NameNode WebUI : Include link to each of datanodes
  • [HADOOP-837] - RunJar should unpack jar files into hadoop.tmp.dir
  • [HADOOP-850] - Add Writable implementations for variable-length integer types.
  • [HADOOP-853] - Move site directories to docs directories

Edit/Copy Release Notes

The text area below allows the project release notes to be edited and copied to another document.