Export - CSV (All fields)
Export - CSV (Current fields)
MAPREDUCE-4981
WordMean, WordMedian, WordStandardDeviation missing from ExamplesDriver
Add toString method to TaggedInputSplit
data_join mavenization broke the mr1 buildMAPREDUCE-4666
JVM metrics for history server
Too many INFO messages written out during AM to RM heartbeatMAPREDUCE-5118
Inverted IndexMAPREDUCE-5116
PUMA Benchmark Suite
JobImpl.finished doesn't expect ERROR as a final job state
Add extra info to JH files
Takes a long time to load the task list on the AM for large jobs
Fix compilation failure of org.apache.hadoop.mapred.gridmix.TestResourceUsageEmulators
AM uses mapreduce.jobtracker.split.metainfo.maxsize but mapred-default has mapreduce.job.split.metainfo.maxsize
Use token request messages defined in hadoop common
TestMRAppMaster#testMRAppMasterMissingStaging occasionally exits
Request with multiple data local nodes can cause NPE in AppSchedulingInfo
"mapred queue -info" spits out ugly exception when queue does not exist
Problem with adding resources when using both -files and -file to hadoop streaming
JobImpl does not handle asynchronous task events in FAILED state
KeyFieldBasedPartitioner not partitioning properly when configured
mr-jobhistory-daemon.sh doesn't accept --config
JobClient.getMapTaskReports on failed job results in NPE
Reducer can fail to make progress during shuffle if too many reducers complete consecutively
Task startup time in JHS is same as job startup time.
DefaultSpeculator generates error messages on normal shutdown
LocalContainerAllocator can loop forever trying to contact the RM
Job End Notification retry interval is 5 milliseconds by default
Reduce MR AM memory usage through String Interning
TestHsWebServicesJobsQuery.testJobsQueryStateInvalid fails with jdk7
AM does not tell the RM about container requests that are no longer needed
Clients should not enforce counter limits
Add experimental support for MR AM to schedule CPUs along-with memory
test-task-controller fails if run as a userid < 1000
mapreduce.user.classpath.first incompatibility with 0.20/1.x
TestYARNRunner.testHistoryServerToken failed on trunk
Type conversion of map completion events leads to performance problems with large jobs
Invalid TaskImpl state transitions when task fails while speculating
MR AM can get in a split brain situation
AM timing out during job commit
Hardcoded task ping timeout kills tasks localizing large amounts of data
ShuffleHandler can generate large logs due to prematurely closed channels
Maven RAT plugin is not checking all source files
JobImpl uber checks for cpu are wrong
Historyserver links expire before the history data does
Speculation + Fetch failures can lead to a hung job
Shuffle race can hang reducer
History files are not move to where the history server sees them
Invalid event: T_ATTEMPT_SUCCEEDED at SUCCEEDED
AM crashes due to OOM while serving up map task completion events