[
https://issues.apache.org/jira/browse/HADOOP-5114?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12669070#action_12669070
]
Raghu Angadi commented on HADOOP-5114:
--------------------------------------
bq. I still want to check why exactly is the timeout for accept() required.
It looks like the timeout for accept is not required (tested on Linux). The
problem was that offerService() thread waits on DataXCeiver thread before
calling shutdown. We should just move the 'join' into shutdown().
will check the simple fix on Windows. Not that TestDiskError fails anyway later
(for a different reason, HADOOP-2023), but I just want to make sure that
DataNode actually exits in case of real error.
> A bunch of mapred unit tests are failing on Windows
> ---------------------------------------------------
>
> Key: HADOOP-5114
> URL: https://issues.apache.org/jira/browse/HADOOP-5114
> Project: Hadoop Core
> Issue Type: Test
> Components: mapred, test
> Affects Versions: 0.18.3
> Environment: Windows
> Reporter: Ramya R
> Assignee: Raghu Angadi
> Priority: Minor
> Fix For: 0.18.4
>
> Attachments: HADOOP-5114.patch, LOGS.zip, tmp.log
>
>
> A bunch of unit tests are consistently failing when run on Windows. Below are
> a list of unit tests which are failing and the corresponding exceptions
> thrown:
> Exception: "java.net.ConnectException: Connection refused: no further
> information"
> Failing tests:
> * TestMiniMRMapRedDebugScript - testMapDebugScript
> * TestNoDefaultsJobConf - testNoDefaults
> * TestQueueManager - testAllEnabledACLForJobSubmission
> * TestCompressedEmptyMapOutputs -
> testMapReduceSortWithCompressedEmptyMapOutputs
> * TestJobInProgressListener - testJobQueueChanges
> * TestKillCompletedJob - testKillCompJob
> * TestMiniMRClasspath - testClassPath
> * TestMiniMRDFSCaching - testWithDFS
> * TestMiniMRWithDFSWithDistinctUsers - testDistinctUsers
> * TestSetupAndCleanupFailure - testWithDFS
> * TestDBJob - testRun
> * TestMiniMRWithDFS - testWithDFS
> * TestJobStatusPersistency - testNonPersistency
> * TestSpecialCharactersInOutputPath - testJobWithDFS
> * TestUserDefinedCounters - testMapReduceJob
> * TestDelegatingInputFormat - testSplitting
> * TestEmptyJobWithDFS - testEmptyJobWithDFS
> * TestJavaSerialization - testMapReduceJob
> * TestClusterMapReduceTestCase - testMapReduce
> Exception: java.lang.IllegalArgumentException: Pathname /<path> from <path>
> is not a valid DFS filename.
> Failing tests:
> * TestJobInProgress - testRunningTaskCount
> * TestJobQueueInformation - testJobQueues
> * TestJobTrackerRestart - testJobTrackerRestart
> Exception: java.io.IOException: Bad connect ack with firstBadLink
> 127.0.0.1:<port number>
> Failing tests:
> * TestJobSysDirWithDFS - testWithDFS
> * TestJobInProgress - testPendingMapTaskCount
> * TestMiniMRDFSSort - testMapReduceSort
> Exception: junit.framework.AssertionFailedError
> Failing tests:
> * TestMRServerPorts - testJobTrackerPorts
> * TestMRServerPorts - testTaskTrackerPorts
> * TestMiniMRTaskTempDir - testTaskTempDir
> Exception: java.io.IOException: Job failed!
> Failing tests:
> * TestMiniMRLocalFS - testWithLocal
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
