[jira] [Moved] (MAPREDUCE-5973) TestAMWebServices* fails intermittently
[ https://issues.apache.org/jira/browse/MAPREDUCE-5973?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Tsuyoshi OZAWA moved YARN-2303 to MAPREDUCE-5973: - Issue Type: Test (was: Bug) Key: MAPREDUCE-5973 (was: YARN-2303) Project: Hadoop Map/Reduce (was: Hadoop YARN) TestAMWebServices* fails intermittently --- Key: MAPREDUCE-5973 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5973 Project: Hadoop Map/Reduce Issue Type: Test Reporter: Tsuyoshi OZAWA Attachments: test-failure-log.txt -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5973) TestAMWebServices* fails intermittently
[ https://issues.apache.org/jira/browse/MAPREDUCE-5973?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Tsuyoshi OZAWA updated MAPREDUCE-5973: -- Description: The tests can fail because of bind exception. TestAMWebServices* fails intermittently --- Key: MAPREDUCE-5973 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5973 Project: Hadoop Map/Reduce Issue Type: Test Reporter: Tsuyoshi OZAWA Attachments: test-failure-log.txt The tests can fail because of bind exception. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5973) TestAMWebServices* fails intermittently
[ https://issues.apache.org/jira/browse/MAPREDUCE-5973?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Tsuyoshi OZAWA updated MAPREDUCE-5973: -- Attachment: test-failure-log.txt TestAMWebServices* fails intermittently --- Key: MAPREDUCE-5973 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5973 Project: Hadoop Map/Reduce Issue Type: Test Reporter: Tsuyoshi OZAWA Attachments: test-failure-log.txt The tests can fail because of bind exception. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5957) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used
[ https://issues.apache.org/jira/browse/MAPREDUCE-5957?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064671#comment-14064671 ] Hadoop QA commented on MAPREDUCE-5957: -- {color:green}+1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12656184/MAPREDUCE-5957.patch against trunk revision . {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 1 new or modified test files. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:green}+1 javadoc{color}. There were no new javadoc warning messages. {color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse. {color:green}+1 findbugs{color}. The patch does not introduce any new Findbugs (version 2.0.3) warnings. {color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings. {color:green}+1 core tests{color}. The patch passed unit tests in hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-common hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient. {color:green}+1 contrib tests{color}. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4746//testReport/ Console output: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4746//console This message is automatically generated. AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used --- Key: MAPREDUCE-5957 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5957 Project: Hadoop Map/Reduce Issue Type: Bug Affects Versions: 2.4.0 Reporter: Sangjin Lee Assignee: Sangjin Lee Attachments: MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch With the job classloader enabled, the MR AM throws ClassNotFoundException if a custom output format class is specified. {noformat} org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:473) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.serviceInit(MRAppMaster.java:374) at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$1.run(MRAppMaster.java:1459) at java.security.AccessController.doPrivileged(Native Method) at javax.security.auth.Subject.doAs(Subject.java:415) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.initAndStartAppMaster(MRAppMaster.java:1456) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.main(MRAppMaster.java:1389) Caused by: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1895) at org.apache.hadoop.mapreduce.task.JobContextImpl.getOutputFormatClass(JobContextImpl.java:222) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:469) ... 8 more Caused by: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1801) at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1893) ... 10 more {noformat} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064692#comment-14064692 ] Rohith commented on MAPREDUCE-5910: --- The test failure is same as MAPREDUCE-5973 MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5971) Move the default options for distcp -p to DistCpOptionSwitch
[ https://issues.apache.org/jira/browse/MAPREDUCE-5971?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064793#comment-14064793 ] Hudson commented on MAPREDUCE-5971: --- FAILURE: Integrated in Hadoop-Yarn-trunk #615 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk/615/]) MAPREDUCE-5971. Move the default options for distcp -p to DistCpOptionSwitch. Contributed by Charles Lamb. (wang: http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1611217) * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-tools/hadoop-distcp/src/main/java/org/apache/hadoop/tools/DistCpOptionSwitch.java * /hadoop/common/trunk/hadoop-tools/hadoop-distcp/src/main/java/org/apache/hadoop/tools/OptionsParser.java Move the default options for distcp -p to DistCpOptionSwitch Key: MAPREDUCE-5971 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5971 Project: Hadoop Map/Reduce Issue Type: Improvement Components: distcp Affects Versions: trunk Reporter: Charles Lamb Assignee: Charles Lamb Priority: Trivial Fix For: 2.6.0 Attachments: MAPREDUCE-5971.001.patch, MAPREDUCE-5971.002.patch The default preserve flags for distcp -p are embedded in the OptionsParser code. Refactor to co-locate them with the actual flag initialization. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5952) LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex
[ https://issues.apache.org/jira/browse/MAPREDUCE-5952?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064792#comment-14064792 ] Hudson commented on MAPREDUCE-5952: --- FAILURE: Integrated in Hadoop-Yarn-trunk #615 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk/615/]) MAPREDUCE-5952. LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex. (Gera Shegalov via kasha) (kasha: http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1611196) * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapred/LocalContainerLauncher.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapred/TestLocalContainerLauncher.java LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex Key: MAPREDUCE-5952 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5952 Project: Hadoop Map/Reduce Issue Type: Bug Components: mr-am, mrv2 Affects Versions: 2.3.0 Reporter: Gera Shegalov Assignee: Gera Shegalov Priority: Blocker Fix For: 2.5.0 Attachments: MAPREDUCE-5952.v01.patch, MAPREDUCE-5952.v02.patch, MAPREDUCE-5952.v03.patch, MAPREDUCE-5952.v04.patch The javadoc comment for {{renameMapOutputForReduce}} incorrectly refers to a single map output directory, whereas this depends on LOCAL_DIRS. mapOutIndex should be set to subMapOutputFile.getOutputIndexFile() {code} 2014-06-30 14:48:35,574 WARN [uber-SubtaskRunner] org.apache.hadoop.mapred.LocalContainerLauncher: Exception running local (uberized) 'child' : java.io.FileNotFoundException: File /Users/gshegalov/workspace/hadoop-common/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/target/org.apache.hadoop.mapreduce.v2.TestMRJobs/org.apache.hadoop.mapreduce.v2. TestMRJobs-localDir-nm-2_3/usercache/gshegalov/appcache/application_1404164272885_0001/output/file.out.index does not exist at org.apache.hadoop.fs.RawLocalFileSystem.deprecatedGetFileStatus(RawLocalFileSystem.java:517) at org.apache.hadoop.fs.RawLocalFileSystem.getFileLinkStatusInternal(RawLocalFileSystem.java:726) at org.apache.hadoop.fs.RawLocalFileSystem.getFileStatus(RawLocalFileSystem.java:507) at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:337) at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:289) at org.apache.hadoop.fs.RawLocalFileSystem.rename(RawLocalFileSystem.java:334) at org.apache.hadoop.fs.ChecksumFileSystem.rename(ChecksumFileSystem.java:504) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.renameMapOutputForReduce(LocalContainerLauncher.java:471) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.runSubtask(LocalContainerLauncher.java:370) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.runTask(LocalContainerLauncher.java:292) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.access$200(LocalContainerLauncher.java:178) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler$1.run(LocalContainerLauncher.java:221) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:895) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:918) at java.lang.Thread.run(Thread.java:695) {code} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-2841) Task level native optimization
[ https://issues.apache.org/jira/browse/MAPREDUCE-2841?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Zhong updated MAPREDUCE-2841: -- Status: Open (was: Patch Available) Task level native optimization -- Key: MAPREDUCE-2841 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2841 Project: Hadoop Map/Reduce Issue Type: Improvement Components: task Environment: x86-64 Linux/Unix Reporter: Binglin Chang Assignee: Sean Zhong Attachments: DESIGN.html, MAPREDUCE-2841.v1.patch, MAPREDUCE-2841.v2.patch, dualpivot-0.patch, dualpivotv20-0.patch, fb-shuffle.patch I'm recently working on native optimization for MapTask based on JNI. The basic idea is that, add a NativeMapOutputCollector to handle k/v pairs emitted by mapper, therefore sort, spill, IFile serialization can all be done in native code, preliminary test(on Xeon E5410, jdk6u24) showed promising results: 1. Sort is about 3x-10x as fast as java(only binary string compare is supported) 2. IFile serialization speed is about 3x of java, about 500MB/s, if hardware CRC32C is used, things can get much faster(1G/ 3. Merge code is not completed yet, so the test use enough io.sort.mb to prevent mid-spill This leads to a total speed up of 2x~3x for the whole MapTask, if IdentityMapper(mapper does nothing) is used There are limitations of course, currently only Text and BytesWritable is supported, and I have not think through many things right now, such as how to support map side combine. I had some discussion with somebody familiar with hive, it seems that these limitations won't be much problem for Hive to benefit from those optimizations, at least. Advices or discussions about improving compatibility are most welcome:) Currently NativeMapOutputCollector has a static method called canEnable(), which checks if key/value type, comparator type, combiner are all compatible, then MapTask can choose to enable NativeMapOutputCollector. This is only a preliminary test, more work need to be done. I expect better final results, and I believe similar optimization can be adopt to reduce task and shuffle too. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-2841) Task level native optimization
[ https://issues.apache.org/jira/browse/MAPREDUCE-2841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064878#comment-14064878 ] Sean Zhong commented on MAPREDUCE-2841: --- Hi Todd, The patch is uploaded to: https://raw.githubusercontent.com/intel-hadoop/nativetask/native_output_collector/patch/hadoop-3.0-mapreduce-2841-2014-7-17.patch (It is too big to be uploaded to here) It is patched against hadoop3.0 trunk. Task level native optimization -- Key: MAPREDUCE-2841 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2841 Project: Hadoop Map/Reduce Issue Type: Improvement Components: task Environment: x86-64 Linux/Unix Reporter: Binglin Chang Assignee: Sean Zhong Attachments: DESIGN.html, MAPREDUCE-2841.v1.patch, MAPREDUCE-2841.v2.patch, dualpivot-0.patch, dualpivotv20-0.patch, fb-shuffle.patch I'm recently working on native optimization for MapTask based on JNI. The basic idea is that, add a NativeMapOutputCollector to handle k/v pairs emitted by mapper, therefore sort, spill, IFile serialization can all be done in native code, preliminary test(on Xeon E5410, jdk6u24) showed promising results: 1. Sort is about 3x-10x as fast as java(only binary string compare is supported) 2. IFile serialization speed is about 3x of java, about 500MB/s, if hardware CRC32C is used, things can get much faster(1G/ 3. Merge code is not completed yet, so the test use enough io.sort.mb to prevent mid-spill This leads to a total speed up of 2x~3x for the whole MapTask, if IdentityMapper(mapper does nothing) is used There are limitations of course, currently only Text and BytesWritable is supported, and I have not think through many things right now, such as how to support map side combine. I had some discussion with somebody familiar with hive, it seems that these limitations won't be much problem for Hive to benefit from those optimizations, at least. Advices or discussions about improving compatibility are most welcome:) Currently NativeMapOutputCollector has a static method called canEnable(), which checks if key/value type, comparator type, combiner are all compatible, then MapTask can choose to enable NativeMapOutputCollector. This is only a preliminary test, more work need to be done. I expect better final results, and I believe similar optimization can be adopt to reduce task and shuffle too. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5952) LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex
[ https://issues.apache.org/jira/browse/MAPREDUCE-5952?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064900#comment-14064900 ] Hudson commented on MAPREDUCE-5952: --- FAILURE: Integrated in Hadoop-Mapreduce-trunk #1834 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/1834/]) MAPREDUCE-5952. LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex. (Gera Shegalov via kasha) (kasha: http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1611196) * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapred/LocalContainerLauncher.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapred/TestLocalContainerLauncher.java LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex Key: MAPREDUCE-5952 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5952 Project: Hadoop Map/Reduce Issue Type: Bug Components: mr-am, mrv2 Affects Versions: 2.3.0 Reporter: Gera Shegalov Assignee: Gera Shegalov Priority: Blocker Fix For: 2.5.0 Attachments: MAPREDUCE-5952.v01.patch, MAPREDUCE-5952.v02.patch, MAPREDUCE-5952.v03.patch, MAPREDUCE-5952.v04.patch The javadoc comment for {{renameMapOutputForReduce}} incorrectly refers to a single map output directory, whereas this depends on LOCAL_DIRS. mapOutIndex should be set to subMapOutputFile.getOutputIndexFile() {code} 2014-06-30 14:48:35,574 WARN [uber-SubtaskRunner] org.apache.hadoop.mapred.LocalContainerLauncher: Exception running local (uberized) 'child' : java.io.FileNotFoundException: File /Users/gshegalov/workspace/hadoop-common/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/target/org.apache.hadoop.mapreduce.v2.TestMRJobs/org.apache.hadoop.mapreduce.v2. TestMRJobs-localDir-nm-2_3/usercache/gshegalov/appcache/application_1404164272885_0001/output/file.out.index does not exist at org.apache.hadoop.fs.RawLocalFileSystem.deprecatedGetFileStatus(RawLocalFileSystem.java:517) at org.apache.hadoop.fs.RawLocalFileSystem.getFileLinkStatusInternal(RawLocalFileSystem.java:726) at org.apache.hadoop.fs.RawLocalFileSystem.getFileStatus(RawLocalFileSystem.java:507) at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:337) at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:289) at org.apache.hadoop.fs.RawLocalFileSystem.rename(RawLocalFileSystem.java:334) at org.apache.hadoop.fs.ChecksumFileSystem.rename(ChecksumFileSystem.java:504) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.renameMapOutputForReduce(LocalContainerLauncher.java:471) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.runSubtask(LocalContainerLauncher.java:370) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.runTask(LocalContainerLauncher.java:292) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.access$200(LocalContainerLauncher.java:178) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler$1.run(LocalContainerLauncher.java:221) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:895) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:918) at java.lang.Thread.run(Thread.java:695) {code} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5971) Move the default options for distcp -p to DistCpOptionSwitch
[ https://issues.apache.org/jira/browse/MAPREDUCE-5971?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064922#comment-14064922 ] Hudson commented on MAPREDUCE-5971: --- FAILURE: Integrated in Hadoop-Hdfs-trunk #1807 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/1807/]) MAPREDUCE-5971. Move the default options for distcp -p to DistCpOptionSwitch. Contributed by Charles Lamb. (wang: http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1611217) * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-tools/hadoop-distcp/src/main/java/org/apache/hadoop/tools/DistCpOptionSwitch.java * /hadoop/common/trunk/hadoop-tools/hadoop-distcp/src/main/java/org/apache/hadoop/tools/OptionsParser.java Move the default options for distcp -p to DistCpOptionSwitch Key: MAPREDUCE-5971 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5971 Project: Hadoop Map/Reduce Issue Type: Improvement Components: distcp Affects Versions: trunk Reporter: Charles Lamb Assignee: Charles Lamb Priority: Trivial Fix For: 2.6.0 Attachments: MAPREDUCE-5971.001.patch, MAPREDUCE-5971.002.patch The default preserve flags for distcp -p are embedded in the OptionsParser code. Refactor to co-locate them with the actual flag initialization. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5952) LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex
[ https://issues.apache.org/jira/browse/MAPREDUCE-5952?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14064921#comment-14064921 ] Hudson commented on MAPREDUCE-5952: --- FAILURE: Integrated in Hadoop-Hdfs-trunk #1807 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/1807/]) MAPREDUCE-5952. LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex. (Gera Shegalov via kasha) (kasha: http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1611196) * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapred/LocalContainerLauncher.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapred/TestLocalContainerLauncher.java LocalContainerLauncher#renameMapOutputForReduce incorrectly assumes a single dir for mapOutIndex Key: MAPREDUCE-5952 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5952 Project: Hadoop Map/Reduce Issue Type: Bug Components: mr-am, mrv2 Affects Versions: 2.3.0 Reporter: Gera Shegalov Assignee: Gera Shegalov Priority: Blocker Fix For: 2.5.0 Attachments: MAPREDUCE-5952.v01.patch, MAPREDUCE-5952.v02.patch, MAPREDUCE-5952.v03.patch, MAPREDUCE-5952.v04.patch The javadoc comment for {{renameMapOutputForReduce}} incorrectly refers to a single map output directory, whereas this depends on LOCAL_DIRS. mapOutIndex should be set to subMapOutputFile.getOutputIndexFile() {code} 2014-06-30 14:48:35,574 WARN [uber-SubtaskRunner] org.apache.hadoop.mapred.LocalContainerLauncher: Exception running local (uberized) 'child' : java.io.FileNotFoundException: File /Users/gshegalov/workspace/hadoop-common/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/target/org.apache.hadoop.mapreduce.v2.TestMRJobs/org.apache.hadoop.mapreduce.v2. TestMRJobs-localDir-nm-2_3/usercache/gshegalov/appcache/application_1404164272885_0001/output/file.out.index does not exist at org.apache.hadoop.fs.RawLocalFileSystem.deprecatedGetFileStatus(RawLocalFileSystem.java:517) at org.apache.hadoop.fs.RawLocalFileSystem.getFileLinkStatusInternal(RawLocalFileSystem.java:726) at org.apache.hadoop.fs.RawLocalFileSystem.getFileStatus(RawLocalFileSystem.java:507) at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:337) at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:289) at org.apache.hadoop.fs.RawLocalFileSystem.rename(RawLocalFileSystem.java:334) at org.apache.hadoop.fs.ChecksumFileSystem.rename(ChecksumFileSystem.java:504) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.renameMapOutputForReduce(LocalContainerLauncher.java:471) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.runSubtask(LocalContainerLauncher.java:370) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.runTask(LocalContainerLauncher.java:292) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler.access$200(LocalContainerLauncher.java:178) at org.apache.hadoop.mapred.LocalContainerLauncher$EventHandler$1.run(LocalContainerLauncher.java:221) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:895) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:918) at java.lang.Thread.run(Thread.java:695) {code} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5957) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used
[ https://issues.apache.org/jira/browse/MAPREDUCE-5957?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sangjin Lee updated MAPREDUCE-5957: --- Status: Open (was: Patch Available) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used --- Key: MAPREDUCE-5957 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5957 Project: Hadoop Map/Reduce Issue Type: Bug Affects Versions: 2.4.0 Reporter: Sangjin Lee Assignee: Sangjin Lee Attachments: MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch With the job classloader enabled, the MR AM throws ClassNotFoundException if a custom output format class is specified. {noformat} org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:473) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.serviceInit(MRAppMaster.java:374) at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$1.run(MRAppMaster.java:1459) at java.security.AccessController.doPrivileged(Native Method) at javax.security.auth.Subject.doAs(Subject.java:415) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.initAndStartAppMaster(MRAppMaster.java:1456) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.main(MRAppMaster.java:1389) Caused by: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1895) at org.apache.hadoop.mapreduce.task.JobContextImpl.getOutputFormatClass(JobContextImpl.java:222) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:469) ... 8 more Caused by: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1801) at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1893) ... 10 more {noformat} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5957) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used
[ https://issues.apache.org/jira/browse/MAPREDUCE-5957?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sangjin Lee updated MAPREDUCE-5957: --- Attachment: MAPREDUCE-5957.patch Fixed the javadoc. AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used --- Key: MAPREDUCE-5957 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5957 Project: Hadoop Map/Reduce Issue Type: Bug Affects Versions: 2.4.0 Reporter: Sangjin Lee Assignee: Sangjin Lee Attachments: MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch With the job classloader enabled, the MR AM throws ClassNotFoundException if a custom output format class is specified. {noformat} org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:473) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.serviceInit(MRAppMaster.java:374) at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$1.run(MRAppMaster.java:1459) at java.security.AccessController.doPrivileged(Native Method) at javax.security.auth.Subject.doAs(Subject.java:415) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.initAndStartAppMaster(MRAppMaster.java:1456) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.main(MRAppMaster.java:1389) Caused by: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1895) at org.apache.hadoop.mapreduce.task.JobContextImpl.getOutputFormatClass(JobContextImpl.java:222) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:469) ... 8 more Caused by: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1801) at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1893) ... 10 more {noformat} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5957) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used
[ https://issues.apache.org/jira/browse/MAPREDUCE-5957?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sangjin Lee updated MAPREDUCE-5957: --- Status: Patch Available (was: Open) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used --- Key: MAPREDUCE-5957 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5957 Project: Hadoop Map/Reduce Issue Type: Bug Affects Versions: 2.4.0 Reporter: Sangjin Lee Assignee: Sangjin Lee Attachments: MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch With the job classloader enabled, the MR AM throws ClassNotFoundException if a custom output format class is specified. {noformat} org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:473) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.serviceInit(MRAppMaster.java:374) at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$1.run(MRAppMaster.java:1459) at java.security.AccessController.doPrivileged(Native Method) at javax.security.auth.Subject.doAs(Subject.java:415) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.initAndStartAppMaster(MRAppMaster.java:1456) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.main(MRAppMaster.java:1389) Caused by: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1895) at org.apache.hadoop.mapreduce.task.JobContextImpl.getOutputFormatClass(JobContextImpl.java:222) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:469) ... 8 more Caused by: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1801) at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1893) ... 10 more {noformat} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-4085) Kill task attempts longer than a configured queue max time
[ https://issues.apache.org/jira/browse/MAPREDUCE-4085?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-4085. - Resolution: Won't Fix Kill task attempts longer than a configured queue max time -- Key: MAPREDUCE-4085 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4085 Project: Hadoop Map/Reduce Issue Type: New Feature Components: task Reporter: Allen Wittenauer Attachments: MAPREDUCE-4085-branch-1.0.4.txt, MAPREDUCE-4085-branch-1.0.txt For some environments, it is desirable to have certain queues have an SLA with regards to task turnover. (i.e., a slot will be free in X minutes and scheduled to the appropriate job) Queues should have a 'task time limit' that would cause task attempts over this time to be killed. This leaves open the possibility that if the task was on a bad node, it could still be rescheduled up to max.task.attempt times. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-4058) adjustable task priority
[ https://issues.apache.org/jira/browse/MAPREDUCE-4058?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-4058. - Resolution: Won't Fix Y! was able to get a different patch into 0.23 and 2.x that provides similar (if limited) capability. adjustable task priority Key: MAPREDUCE-4058 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4058 Project: Hadoop Map/Reduce Issue Type: New Feature Components: task-controller Affects Versions: 1.0.0 Reporter: Allen Wittenauer Assignee: Mark Wagner Attachments: MAPREDUCE-4058-branch-1.0.patch For those of us that completely destroy our CPUs, it is beneficial to be able to run user tasks at a different priority than the tasktracker. This would allow for TTs (and by extension, DNs) to get more CPU clock cycles so that things like heartbeats don't disappear. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-2378) Reduce fails when running on 1 small file.
[ https://issues.apache.org/jira/browse/MAPREDUCE-2378?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-2378. - Resolution: Cannot Reproduce Closing this as 'cannot reproduce' as log4j has since been upgraded. A few times, actually. Reduce fails when running on 1 small file. --- Key: MAPREDUCE-2378 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2378 Project: Hadoop Map/Reduce Issue Type: Bug Affects Versions: 0.21.0 Environment: java version 1.6.0_07 Diablo Java(TM) SE Runtime Environment (build 1.6.0_07-b02) Diablo Java HotSpot(TM) 64-Bit Server VM (build 10.0-b23, mixed mode) Reporter: Simon Dircks Labels: 1, failed, file, log4j, reduce, single, small, tiny Attachments: failed reduce task log.html If i run the wordcount example on 1 small (less than 2MB) file i get the following error: log4j:ERROR Failed to flush writer, java.io.InterruptedIOException at java.io.FileOutputStream.writeBytes(Native Method) at java.io.FileOutputStream.write(FileOutputStream.java:260) at sun.nio.cs.StreamEncoder.writeBytes(StreamEncoder.java:202) at sun.nio.cs.StreamEncoder.implFlushBuffer(StreamEncoder.java:272) at sun.nio.cs.StreamEncoder.implFlush(StreamEncoder.java:276) at sun.nio.cs.StreamEncoder.flush(StreamEncoder.java:122) at java.io.OutputStreamWriter.flush(OutputStreamWriter.java:212) at org.apache.log4j.helpers.QuietWriter.flush(QuietWriter.java:58) at org.apache.log4j.WriterAppender.subAppend(WriterAppender.java:316) at org.apache.log4j.WriterAppender.append(WriterAppender.java:160) at org.apache.hadoop.mapred.TaskLogAppender.append(TaskLogAppender.java:58) at org.apache.log4j.AppenderSkeleton.doAppend(AppenderSkeleton.java:251) at org.apache.log4j.helpers.AppenderAttachableImpl.appendLoopOnAppenders(AppenderAttachableImpl.java:66) at org.apache.log4j.Category.callAppenders(Category.java:206) at org.apache.log4j.Category.forcedLog(Category.java:391) at org.apache.log4j.Category.log(Category.java:856) at org.apache.commons.logging.impl.Log4JLogger.info(Log4JLogger.java:199) at org.apache.hadoop.mapreduce.task.reduce.ShuffleScheduler.freeHost(ShuffleScheduler.java:345) at org.apache.hadoop.mapreduce.task.reduce.Fetcher.run(Fetcher.java:152) If i run the wordcount test with 2 files, it works fine. I have actually repeated this with my own code. I am working on something that requires me to map/reduce a small file and I had to work around the problem by splitting the file into 2 1MB pieces for my job to run. All our jobs that run on 1 single larger file (over 1GB) work flawlessly. I am not exactly sure the threshold, From the testing i have done it seems to be any file smaller than the default HDFS block size (64MB) Sometimes it seems random in the 5-64MB range. But its 100% for the 5MB and smaller files. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-2444) error connect tasktracker for jobtracker
[ https://issues.apache.org/jira/browse/MAPREDUCE-2444?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-2444. - Resolution: Duplicate This was essentially resolved with the move to Protobuf for the RPC layer. error connect tasktracker for jobtracker Key: MAPREDUCE-2444 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2444 Project: Hadoop Map/Reduce Issue Type: Bug Components: tasktracker Affects Versions: 0.20.1 Reporter: Alexey Diomin Labels: hadoop In TaskTracker.java on create connection to JobTracker we check compare build version if(!VersionInfo.getBuildVersion().equals(jobTrackerBV)) but public static String getBuildVersion(){ return VersionInfo.getVersion() + from + VersionInfo.getRevision() + by + VersionInfo.getUser() + source checksum + VersionInfo.getSrcChecksum(); } in result identical version/revision/srcChecksum but compiled different users incompatible -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-2841) Task level native optimization
[ https://issues.apache.org/jira/browse/MAPREDUCE-2841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065042#comment-14065042 ] Todd Lipcon commented on MAPREDUCE-2841: Hey Sean. Something seems to be wrong with that patchfile -- some of the files seem to be present 11 times in it: {code} todd@todd-ThinkPad-T540p:~$ grep '+++.*TextSerializer' hadoop-3.0-mapreduce-2841-2014-7-17.patch | less -S | wc -l 11 {code} That might also explain why the file is too large to upload here as an attachment. Could you try to regenerate the patch? Task level native optimization -- Key: MAPREDUCE-2841 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2841 Project: Hadoop Map/Reduce Issue Type: Improvement Components: task Environment: x86-64 Linux/Unix Reporter: Binglin Chang Assignee: Sean Zhong Attachments: DESIGN.html, MAPREDUCE-2841.v1.patch, MAPREDUCE-2841.v2.patch, dualpivot-0.patch, dualpivotv20-0.patch, fb-shuffle.patch I'm recently working on native optimization for MapTask based on JNI. The basic idea is that, add a NativeMapOutputCollector to handle k/v pairs emitted by mapper, therefore sort, spill, IFile serialization can all be done in native code, preliminary test(on Xeon E5410, jdk6u24) showed promising results: 1. Sort is about 3x-10x as fast as java(only binary string compare is supported) 2. IFile serialization speed is about 3x of java, about 500MB/s, if hardware CRC32C is used, things can get much faster(1G/ 3. Merge code is not completed yet, so the test use enough io.sort.mb to prevent mid-spill This leads to a total speed up of 2x~3x for the whole MapTask, if IdentityMapper(mapper does nothing) is used There are limitations of course, currently only Text and BytesWritable is supported, and I have not think through many things right now, such as how to support map side combine. I had some discussion with somebody familiar with hive, it seems that these limitations won't be much problem for Hive to benefit from those optimizations, at least. Advices or discussions about improving compatibility are most welcome:) Currently NativeMapOutputCollector has a static method called canEnable(), which checks if key/value type, comparator type, combiner are all compatible, then MapTask can choose to enable NativeMapOutputCollector. This is only a preliminary test, more work need to be done. I expect better final results, and I believe similar optimization can be adopt to reduce task and shuffle too. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-2841) Task level native optimization
[ https://issues.apache.org/jira/browse/MAPREDUCE-2841?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Zhong updated MAPREDUCE-2841: -- Attachment: hadoop-3.0-mapreduce-2841-2014-7-17.patch Task level native optimization -- Key: MAPREDUCE-2841 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2841 Project: Hadoop Map/Reduce Issue Type: Improvement Components: task Environment: x86-64 Linux/Unix Reporter: Binglin Chang Assignee: Sean Zhong Attachments: DESIGN.html, MAPREDUCE-2841.v1.patch, MAPREDUCE-2841.v2.patch, dualpivot-0.patch, dualpivotv20-0.patch, fb-shuffle.patch, hadoop-3.0-mapreduce-2841-2014-7-17.patch I'm recently working on native optimization for MapTask based on JNI. The basic idea is that, add a NativeMapOutputCollector to handle k/v pairs emitted by mapper, therefore sort, spill, IFile serialization can all be done in native code, preliminary test(on Xeon E5410, jdk6u24) showed promising results: 1. Sort is about 3x-10x as fast as java(only binary string compare is supported) 2. IFile serialization speed is about 3x of java, about 500MB/s, if hardware CRC32C is used, things can get much faster(1G/ 3. Merge code is not completed yet, so the test use enough io.sort.mb to prevent mid-spill This leads to a total speed up of 2x~3x for the whole MapTask, if IdentityMapper(mapper does nothing) is used There are limitations of course, currently only Text and BytesWritable is supported, and I have not think through many things right now, such as how to support map side combine. I had some discussion with somebody familiar with hive, it seems that these limitations won't be much problem for Hive to benefit from those optimizations, at least. Advices or discussions about improving compatibility are most welcome:) Currently NativeMapOutputCollector has a static method called canEnable(), which checks if key/value type, comparator type, combiner are all compatible, then MapTask can choose to enable NativeMapOutputCollector. This is only a preliminary test, more work need to be done. I expect better final results, and I believe similar optimization can be adopt to reduce task and shuffle too. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Created] (MAPREDUCE-5974) Allow map output collector fallback
Todd Lipcon created MAPREDUCE-5974: -- Summary: Allow map output collector fallback Key: MAPREDUCE-5974 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5974 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Affects Versions: 2.6.0 Reporter: Todd Lipcon Currently we only allow specifying a single MapOutputCollector implementation class in a job. It would be nice to allow a comma-separated list of classes: we should try each collector implementation in the user-specified order until we find one that can be successfully instantiated and initted. This is useful for cases where a particular optimized collector implementation cannot operate on all key/value types, or requires native code. The cluster administrator can configure the cluster to try to use the optimized collector and fall back to the default collector. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-2841) Task level native optimization
[ https://issues.apache.org/jira/browse/MAPREDUCE-2841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065064#comment-14065064 ] Sean Zhong commented on MAPREDUCE-2841: --- Ah, thanks for pointing this out. I am not sure why this happen. I just uploaded the patch to this jira https://issues.apache.org/jira/secure/attachment/12656288/hadoop-3.0-mapreduce-2841-2014-7-17.patch updates: 1. Remove Hbase/hive/hive/mahout/pig related code, those code will be posted elsewhere in another jira or hosted on github. 2. Use ServiceLoader to discover custom platform(to support custom key types) Task level native optimization -- Key: MAPREDUCE-2841 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2841 Project: Hadoop Map/Reduce Issue Type: Improvement Components: task Environment: x86-64 Linux/Unix Reporter: Binglin Chang Assignee: Sean Zhong Attachments: DESIGN.html, MAPREDUCE-2841.v1.patch, MAPREDUCE-2841.v2.patch, dualpivot-0.patch, dualpivotv20-0.patch, fb-shuffle.patch, hadoop-3.0-mapreduce-2841-2014-7-17.patch I'm recently working on native optimization for MapTask based on JNI. The basic idea is that, add a NativeMapOutputCollector to handle k/v pairs emitted by mapper, therefore sort, spill, IFile serialization can all be done in native code, preliminary test(on Xeon E5410, jdk6u24) showed promising results: 1. Sort is about 3x-10x as fast as java(only binary string compare is supported) 2. IFile serialization speed is about 3x of java, about 500MB/s, if hardware CRC32C is used, things can get much faster(1G/ 3. Merge code is not completed yet, so the test use enough io.sort.mb to prevent mid-spill This leads to a total speed up of 2x~3x for the whole MapTask, if IdentityMapper(mapper does nothing) is used There are limitations of course, currently only Text and BytesWritable is supported, and I have not think through many things right now, such as how to support map side combine. I had some discussion with somebody familiar with hive, it seems that these limitations won't be much problem for Hive to benefit from those optimizations, at least. Advices or discussions about improving compatibility are most welcome:) Currently NativeMapOutputCollector has a static method called canEnable(), which checks if key/value type, comparator type, combiner are all compatible, then MapTask can choose to enable NativeMapOutputCollector. This is only a preliminary test, more work need to be done. I expect better final results, and I believe similar optimization can be adopt to reduce task and shuffle too. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5957) AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used
[ https://issues.apache.org/jira/browse/MAPREDUCE-5957?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065068#comment-14065068 ] Hadoop QA commented on MAPREDUCE-5957: -- {color:green}+1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12656274/MAPREDUCE-5957.patch against trunk revision . {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 1 new or modified test files. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:green}+1 javadoc{color}. There were no new javadoc warning messages. {color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse. {color:green}+1 findbugs{color}. The patch does not introduce any new Findbugs (version 2.0.3) warnings. {color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings. {color:green}+1 core tests{color}. The patch passed unit tests in hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-common hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient. {color:green}+1 contrib tests{color}. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4747//testReport/ Console output: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4747//console This message is automatically generated. AM throws ClassNotFoundException with job classloader enabled if custom output format/committer is used --- Key: MAPREDUCE-5957 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5957 Project: Hadoop Map/Reduce Issue Type: Bug Affects Versions: 2.4.0 Reporter: Sangjin Lee Assignee: Sangjin Lee Attachments: MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch, MAPREDUCE-5957.patch With the job classloader enabled, the MR AM throws ClassNotFoundException if a custom output format class is specified. {noformat} org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:473) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.serviceInit(MRAppMaster.java:374) at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$1.run(MRAppMaster.java:1459) at java.security.AccessController.doPrivileged(Native Method) at javax.security.auth.Subject.doAs(Subject.java:415) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.initAndStartAppMaster(MRAppMaster.java:1456) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.main(MRAppMaster.java:1389) Caused by: java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1895) at org.apache.hadoop.mapreduce.task.JobContextImpl.getOutputFormatClass(JobContextImpl.java:222) at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.createOutputCommitter(MRAppMaster.java:469) ... 8 more Caused by: java.lang.ClassNotFoundException: Class com.foo.test.TestOutputFormat not found at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1801) at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1893) ... 10 more {noformat} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-296) job statistics should be displayed in the web/ui
[ https://issues.apache.org/jira/browse/MAPREDUCE-296?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-296. Resolution: Duplicate This was done back in 0.20 as part of the y! security merge. job statistics should be displayed in the web/ui Key: MAPREDUCE-296 URL: https://issues.apache.org/jira/browse/MAPREDUCE-296 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Owen O'Malley Labels: newbie It would be really nice, if the job page in the web/ui showed the time that: 1. first map started 2. last map finished 3. last reduce finished shuffle 4. last reduce finished sort 5. last reduce finished -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-603) Fix unchecked warnings in contrib code
[ https://issues.apache.org/jira/browse/MAPREDUCE-603?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-603. Resolution: Unresolved Stale. Closing this out. Fix unchecked warnings in contrib code -- Key: MAPREDUCE-603 URL: https://issues.apache.org/jira/browse/MAPREDUCE-603 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: Tom White There are unchecked warnings in abacus, data_join and streaming. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-135) speculative task failure can kill jobs
[ https://issues.apache.org/jira/browse/MAPREDUCE-135?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-135. Resolution: Incomplete I'm going to close this out as stale. I suspect this is no longer an issue. speculative task failure can kill jobs -- Key: MAPREDUCE-135 URL: https://issues.apache.org/jira/browse/MAPREDUCE-135 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Owen O'Malley We had a case where the random writer example was killed by speculative execution. It happened like: task_0001_m_000123_0 - starts task_0001_m_000123_1 - starts and fails because attempt 0 is creating the file task_0001_m_000123_2 - starts and fails because attempt 0 is creating the file task_0001_m_000123_3 - starts and fails because attempt 0 is creating the file task_0001_m_000123_4 - starts and fails because attempt 0 is creating the file job_0001 is killed because map_000123 failed 4 times. From this experience, I think we should change the scheduling so that: 1. Tasks are only allowed 1 speculative attempt. 2. TIPs don't kill jobs until they have 4 failures AND the last task under that tip fails. Thoughts? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-585) A corrupt text file causes the maps to hang
[ https://issues.apache.org/jira/browse/MAPREDUCE-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-585. Resolution: Incomplete I'm going to close this out as stale. I suspect this is no longer an issue. A corrupt text file causes the maps to hang --- Key: MAPREDUCE-585 URL: https://issues.apache.org/jira/browse/MAPREDUCE-585 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: Mahadev konar Priority: Minor A corrupt file hangs a map. The map keeps reading the same record again and again and never finishes. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-167) SAXParseException causes test to run forever
[ https://issues.apache.org/jira/browse/MAPREDUCE-167?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-167. Resolution: Incomplete I'm going to close this out as stale. I suspect this is no longer an issue. SAXParseException causes test to run forever Key: MAPREDUCE-167 URL: https://issues.apache.org/jira/browse/MAPREDUCE-167 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Nigel Daley Attachments: thread.dump.txt Occassionally, while running TestMiniMRClasspath, I get a SAXParseException that causes the test to run forever. Two questions I have: 1) what is the underlying cause of the SAXParseException? 2) does the JobTracker realize that a task got lost? Here's the pertinent test trace: [junit] 2007-02-13 19:26:56,058 INFO mapred.JobClient (JobClient.java:runJob(400)) - Running job: job_0001 [junit] 2007-02-13 19:26:57,062 INFO mapred.JobClient (JobClient.java:runJob(417)) - map 0% reduce 0% [junit] 2007-02-13 19:27:05,258 INFO mapred.JobInProgress (JobInProgress.java:findNewTask(421)) - Choosing cached task tip_0001_m_00 [junit] 2007-02-13 19:27:05,259 INFO mapred.JobTracker (JobTracker.java:createTaskEntry(690)) - Adding task 'task_0001_m_00_0' to tip tip_0001_m_00, for tracker 'tracker_ucdev15.yst.corp.yahoo.com:50067' [junit] 2007-02-13 19:27:05,260 INFO mapred.JobInProgress (JobInProgress.java:findNewTask(421)) - Choosing cached task tip_0001_m_01 [junit] 2007-02-13 19:27:05,261 INFO mapred.JobTracker (JobTracker.java:createTaskEntry(690)) - Adding task 'task_0001_m_01_0' to tip tip_0001_m_01, for tracker 'tracker_ucdev15.yst.corp.yahoo.com:50063' [junit] 2007-02-13 19:27:05,262 INFO mapred.TaskTracker (TaskTracker.java:startNewTask(822)) - LaunchTaskAction: task_0001_m_00_0 [junit] 2007-02-13 19:27:05,262 INFO mapred.JobInProgress (JobInProgress.java:findNewTask(421)) - Choosing cached task tip_0001_m_02 [junit] 2007-02-13 19:27:05,263 INFO mapred.JobTracker (JobTracker.java:createTaskEntry(690)) - Adding task 'task_0001_m_02_0' to tip tip_0001_m_02, for tracker 'tracker_ucdev15.yst.corp.yahoo.com:50066' [junit] 2007-02-13 19:27:05,263 INFO mapred.TaskTracker (TaskTracker.java:startNewTask(822)) - LaunchTaskAction: task_0001_m_01_0 [junit] 2007-02-13 19:27:05,267 INFO mapred.TaskTracker (TaskTracker.java:startNewTask(822)) - LaunchTaskAction: task_0001_m_02_0 [junit] 2007-02-13 19:27:05,270 INFO mapred.JobInProgress (JobInProgress.java:findNewTask(453)) - Choosing normal task tip_0001_r_00 [junit] 2007-02-13 19:27:05,270 INFO mapred.JobTracker (JobTracker.java:createTaskEntry(690)) - Adding task 'task_0001_r_00_0' to tip tip_0001_r_00, for tracker 'tracker_ucdev15.yst.corp.yahoo.com:50062' [junit] 2007-02-13 19:27:05,271 INFO mapred.TaskTracker (TaskTracker.java:startNewTask(822)) - LaunchTaskAction: task_0001_r_00_0 [junit] 2007-02-13 19:27:05,285 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_-4805938806139473507 to /66.228.166.95 [junit] 2007-02-13 19:27:05,289 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_-4805938806139473507 to /66.228.166.95 [junit] 2007-02-13 19:27:05,292 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_-4805938806139473507 to /66.228.166.95 [junit] 2007-02-13 19:27:05,295 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_-4805938806139473507 to /66.228.166.95 [junit] 2007-02-13 19:27:05,312 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_3019208026182045172 to /66.228.166.95 [junit] 2007-02-13 19:27:05,312 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_3019208026182045172 to /66.228.166.95 [junit] 2007-02-13 19:27:05,352 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_-1390246588917827761 to /66.228.166.95 [junit] 2007-02-13 19:27:05,355 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_-1390246588917827761 to /66.228.166.95 [junit] 2007-02-13 19:27:05,367 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_4739954315939188869 to /66.228.166.95 [junit] 2007-02-13 19:27:05,368 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_3019208026182045172 to /66.228.166.95 [junit] 2007-02-13 19:27:05,367 INFO dfs.DataNode (DataNode.java:readBlock(719)) - Served block blk_4739954315939188869 to /66.228.166.95 [junit] 2007-02-13 19:27:05,416 FATAL conf.Configuration (Configuration.java:loadResource(552)) - error parsing conf file:
[jira] [Resolved] (MAPREDUCE-584) In Streaming, crashes after all the input is consumed, are not detected
[ https://issues.apache.org/jira/browse/MAPREDUCE-584?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-584. Resolution: Fixed I'm going to close this out as stale. I suspect this is no longer an issue. In Streaming, crashes after all the input is consumed, are not detected --- Key: MAPREDUCE-584 URL: https://issues.apache.org/jira/browse/MAPREDUCE-584 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: arkady borkovsky Priority: Minor In a Hadoop Streaming, if the user code crashes after all the input has been consumed, the framework considers the process to be successful. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-306) job submission protocol should have a method for getting the task capacity of the cluster
[ https://issues.apache.org/jira/browse/MAPREDUCE-306?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-306. Resolution: Fixed Now that we have more experience, it is generally recognized that using all of the map slots for a job is a terrible idea. Closing. job submission protocol should have a method for getting the task capacity of the cluster --- Key: MAPREDUCE-306 URL: https://issues.apache.org/jira/browse/MAPREDUCE-306 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Owen O'Malley It would help the InputFormats make informed decisions if the JobSubmissionProtocol had a method the returned the number of tasks that the cluster can run at once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-191) Mapper fail rate increases significantly as the number of reduces increase
[ https://issues.apache.org/jira/browse/MAPREDUCE-191?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-191. Resolution: Cannot Reproduce I'm going to close this out as stale. I suspect this is no longer an issue. Mapper fail rate increases significantly as the number of reduces increase -- Key: MAPREDUCE-191 URL: https://issues.apache.org/jira/browse/MAPREDUCE-191 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Runping Qi I ran a large sort job, with about 8400 mappers. In the first run, I used 301 reducers. About 600 mapper tasks failed. In another run, I used 607 reducers. More than 3800 mapper tasks failed. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-172) Reducers stuck in 'sort'
[ https://issues.apache.org/jira/browse/MAPREDUCE-172?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-172. Resolution: Fixed I'm going to close this out as stale. I suspect this is no longer an issue. Reducers stuck in 'sort' Key: MAPREDUCE-172 URL: https://issues.apache.org/jira/browse/MAPREDUCE-172 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Arun C Murthy A couple of reduces seem stuck on a small 20-node cluster in the 'sort' phase for almost an hour: TaskTracker logs: 2007-03-28 14:13:46,471 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_05_0 0.3334% reduce sort 2007-03-28 14:13:46,478 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_09_0 0.3334% reduce sort 2007-03-28 14:13:47,476 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_05_0 0.3334% reduce sort 2007-03-28 14:13:47,483 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_09_0 0.3334% reduce sort ... ... ... 2007-03-28 15:06:04,376 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_05_0 0.3334% reduce sort 2007-03-28 15:06:04,411 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_09_0 0.3334% reduce sort 2007-03-28 15:06:05,379 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_05_0 0.3334% reduce sort 2007-03-28 15:06:05,414 INFO org.apache.hadoop.mapred.TaskTracker: task_0002_r_09_0 0.3334% reduce sort Eventually the JobTracker declared the same TT 'lost' (presumably for no heartbeats): 2007-03-28 15:18:20,341 INFO org.apache.hadoop.mapred.JobTracker: Lost tracker 'tracker_XXX:9020' -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-602) The streaming code should be moved from contrib to Hadoop main framework
[ https://issues.apache.org/jira/browse/MAPREDUCE-602?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-602. Resolution: Fixed This happened forever ago. The streaming code should be moved from contrib to Hadoop main framework Key: MAPREDUCE-602 URL: https://issues.apache.org/jira/browse/MAPREDUCE-602 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: Runping Qi Before the actual move, the code needs a bit of further clean up in the following areas: 1. coding style/convention, and code quality 2. XMLRecordReader: the current implementation is too hacky. 3. Better javadoc -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-417) Logging could hang/fail when drive is filled by mapred outputs.
[ https://issues.apache.org/jira/browse/MAPREDUCE-417?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-417. Resolution: Fixed I'm going to mark this as fixed due to how user logging is now handled. Logging could hang/fail when drive is filled by mapred outputs. --- Key: MAPREDUCE-417 URL: https://issues.apache.org/jira/browse/MAPREDUCE-417 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Koji Noguchi Priority: Minor HADOOP-1252 addresses the mapred disk problems. In addition to those problems, if mapred fills up the drive used for logging, it might affect TaskTracker/DataNodes. Simple solution for now is not to use the logging drive in MapReduce. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-452) tasktracker checkpointing capability
[ https://issues.apache.org/jira/browse/MAPREDUCE-452?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-452. Resolution: Fixed Marking this as fixed since YARN provides this capability. tasktracker checkpointing capability Key: MAPREDUCE-452 URL: https://issues.apache.org/jira/browse/MAPREDUCE-452 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Pete Wyckoff Priority: Minor This relates to allowing a resource manager (e.g., hadoop on demand) to grow and (rarely) shrink jobs on the fly. Growing is already supported. Shrinking could be done in 2 ways - (1) consider the machine dead and allow speculative execution to take care of it or (2) moving the existing map outputs from that machine somewhere else (another machine, dfs) - task tracker checkpointing In the case of IO only intensive jobs, checkpointing the tasktracker doesn't do much for you. But, in the case of CPU or other scarce resource (e.g., a DB or Webpage cache...), the checkpointing could be very useful. The question is how often is this the case and how useful? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-107) Tasks fail due to lost mapout
[ https://issues.apache.org/jira/browse/MAPREDUCE-107?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-107. Resolution: Fixed I'm going to close this out as stale. I suspect this is no longer an issue. Tasks fail due to lost mapout - Key: MAPREDUCE-107 URL: https://issues.apache.org/jira/browse/MAPREDUCE-107 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Hairong Kuang When I ran a job, each map task of which generates Gbytes of map output, I saw many tasks failed with following errors: Map output lost, rescheduling: getMapOutput(task_0993_m_13_0,140) failed : java.io.FileNotFoundException: /hadoop/mapred/local/task_0993_m_13_0/file.out at org.apache.hadoop.fs.ChecksumFileSystem.open(ChecksumFileSystem.java:332) at org.apache.hadoop.fs.FileSystem.open(FileSystem.java:245) at org.apache.hadoop.mapred.TaskTracker$MapOutputServlet.doGet(TaskTracker.java:1657) at javax.servlet.http.HttpServlet.service(HttpServlet.java:689) at javax.servlet.http.HttpServlet.service(HttpServlet.java:802) at org.mortbay.jetty.servlet.ServletHolder.handle(ServletHolder.java:427) at org.mortbay.jetty.servlet.WebApplicationHandler.dispatch(WebApplicationHandler.java:475) at org.mortbay.jetty.servlet.ServletHandler.handle(ServletHandler.java:567) at org.mortbay.http.HttpContext.handle(HttpContext.java:1565) at org.mortbay.jetty.servlet.WebApplicationContext.handle(WebApplicationContext.java:635) at org.mortbay.http.HttpContext.handle(HttpContext.java:1517) at org.mortbay.http.HttpServer.service(HttpServer.java:954) at org.mortbay.http.HttpConnection.service(HttpConnection.java:814) at org.mortbay.http.HttpConnection.handleNext(HttpConnection.java:981) at org.mortbay.http.HttpConnection.handle(HttpConnection.java:831) at org.mortbay.http.SocketListener.handleConnection(SocketListener.java:244) at org.mortbay.util.ThreadedServer.handle(ThreadedServer.java:357) at org.mortbay.util.ThreadPool$PoolThread.run(ThreadPool.java:534) -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-319) Use Grizzly for Fetching Map Output in Shuffle
[ https://issues.apache.org/jira/browse/MAPREDUCE-319?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-319. Resolution: Fixed Yes. Closing. Use Grizzly for Fetching Map Output in Shuffle -- Key: MAPREDUCE-319 URL: https://issues.apache.org/jira/browse/MAPREDUCE-319 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Tahir Hashmi Assignee: Devaraj Das Attachments: 1432.patch, grizzly.tgz As mentioned in HADOOP-1273 and references therefrom, Jetty 6 still doesn't seem to be stable enough for use in Hadoop. Instead, we've decided to consider the usage of Grizzly Framework [https://grizzly.dev.java.net/] for NIO based communication. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-316) Splittability of input should be controllable by application
[ https://issues.apache.org/jira/browse/MAPREDUCE-316?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-316. Resolution: Won't Fix Closing this as won't fix. As was pointed out, there are other ways. Splittability of input should be controllable by application Key: MAPREDUCE-316 URL: https://issues.apache.org/jira/browse/MAPREDUCE-316 Project: Hadoop Map/Reduce Issue Type: Improvement Environment: ALL Reporter: Milind Bhandarkar Assignee: Senthil Subramanian Attachments: HADOOP-1441_1.patch Currently, isSplittable method of FileInputFormat always returns true. For some applications, it becomes necessary that the map task process entire file, rather than a block. Therefore, splittability of input (i.e. block-level split vs file-level-split) should be controllable by user via a configuration variable. The default could be block-level split, as is. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-320) Map/Reduce should use IP addresses to identify nodes rather than hostnames
[ https://issues.apache.org/jira/browse/MAPREDUCE-320?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065159#comment-14065159 ] Allen Wittenauer commented on MAPREDUCE-320: I'm tempted to close this as won't fix. The fundamental problem is that the MR framework (and all other Hadoop systems) need to a way to distinguish a process as the same host on machines with multiple interfaces. This is certainly fixable... and the various token systems we have floating around may actually do that. But it's a tremendous amount of work (especially from a security perspective) and I don't see much interest in doing that work. Map/Reduce should use IP addresses to identify nodes rather than hostnames -- Key: MAPREDUCE-320 URL: https://issues.apache.org/jira/browse/MAPREDUCE-320 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Owen O'Malley Assignee: Owen O'Malley We should move the Map/Reduce framework to identify hosts as IP addresses rather than hostnames to prevent problems with DNS. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Assigned] (MAPREDUCE-5974) Allow map output collector fallback
[ https://issues.apache.org/jira/browse/MAPREDUCE-5974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Todd Lipcon reassigned MAPREDUCE-5974: -- Assignee: Todd Lipcon Allow map output collector fallback --- Key: MAPREDUCE-5974 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5974 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Affects Versions: 2.6.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Attachments: mapreduce-5974.txt Currently we only allow specifying a single MapOutputCollector implementation class in a job. It would be nice to allow a comma-separated list of classes: we should try each collector implementation in the user-specified order until we find one that can be successfully instantiated and initted. This is useful for cases where a particular optimized collector implementation cannot operate on all key/value types, or requires native code. The cluster administrator can configure the cluster to try to use the optimized collector and fall back to the default collector. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5974) Allow map output collector fallback
[ https://issues.apache.org/jira/browse/MAPREDUCE-5974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Todd Lipcon updated MAPREDUCE-5974: --- Attachment: mapreduce-5974.txt Attached patch implements the improvement as described. I did not include any new unit test since this code path is exercised by existing paths, and mocking out this section of the code is really quite difficult. If folks would like to see a unit test, I can add a full functional test which specifies some bogus collector class ahead of the real implementation, but figured that it's a trivial enough change we could get by without. Allow map output collector fallback --- Key: MAPREDUCE-5974 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5974 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Affects Versions: 2.6.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Attachments: mapreduce-5974.txt Currently we only allow specifying a single MapOutputCollector implementation class in a job. It would be nice to allow a comma-separated list of classes: we should try each collector implementation in the user-specified order until we find one that can be successfully instantiated and initted. This is useful for cases where a particular optimized collector implementation cannot operate on all key/value types, or requires native code. The cluster administrator can configure the cluster to try to use the optimized collector and fall back to the default collector. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-1766) Hadoop Streaming should not use TextInputFormat class as the default input format class.
[ https://issues.apache.org/jira/browse/MAPREDUCE-1766?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-1766. - Resolution: Fixed Closing as stale, especially given that one can set the inputformat for streaming jobs now. Hadoop Streaming should not use TextInputFormat class as the default input format class. Key: MAPREDUCE-1766 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1766 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: Runping Qi The TextInputFormat class does not work with IdentityMapper class. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated MAPREDUCE-5910: --- Status: Open (was: Patch Available) MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated MAPREDUCE-5910: --- Attachment: MAPREDUCE-5910.4.patch I see, thanks for investigating. added one code comment myself, re-submit the patch. MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch, MAPREDUCE-5910.4.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated MAPREDUCE-5910: --- Status: Patch Available (was: Open) MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch, MAPREDUCE-5910.4.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-259) Rack-aware Shuffle
[ https://issues.apache.org/jira/browse/MAPREDUCE-259?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-259. Resolution: Duplicate Rack-aware Shuffle -- Key: MAPREDUCE-259 URL: https://issues.apache.org/jira/browse/MAPREDUCE-259 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Arun C Murthy Assignee: Arun C Murthy We could try and experiment with *rack-aware* scheduling of fetches per-reducer. Given the disparities between in-rack and off-rack bandwidth it could be a improvement to do something along these lines: {noformat} if (no. of known map-output locations than no. of copier threads) { try to schedule 75% of copies off-rack try schedule 25% of copies in-rack } {noformat} This could lead to better utilization of both in-rack switch b/w... Clearly we want to schedule more cross-switch than in-rack since off-rack copies will take significantly more time; hence the 75-25 split. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-426) Race condition in LaunchTaskAction and KillJobAction
[ https://issues.apache.org/jira/browse/MAPREDUCE-426?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-426. Resolution: Fixed I'm going to close this as a stale issue. There have been a lot of race conditions fixed in this area and suspect this is one of them. Race condition in LaunchTaskAction and KillJobAction Key: MAPREDUCE-426 URL: https://issues.apache.org/jira/browse/MAPREDUCE-426 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Koji Noguchi Priority: Minor One task wasn't killed when its job was killed. On the TaskTracker log, it showed, 2007-08-21 17:02:29,219 INFO org.apache.hadoop.mapred.TaskTracker: LaunchTaskAction: task_0133_r_80_2** 2007-08-21 17:02:29,232 INFO org.apache.hadoop.mapred.TaskTracker: Received 'KillJobAction' for job: job_0131 ** 2007-08-21 17:02:29,233 INFO org.apache.hadoop.mapred.TaskRunner: task_0131_m_77_0 done; removing files. 2007-08-21 17:02:29,376 INFO org.apache.hadoop.mapred.TaskTracker: Received 'KillJobAction' for job: job_0133 2007-08-21 17:02:29,376 INFO org.apache.hadoop.mapred.TaskRunner: task_0133_r_60_0 done; removing files. 2007-08-21 17:02:29,378 INFO org.apache.hadoop.mapred.TaskRunner: task_0133_r_71_2 done; removing files. 2007-08-21 17:02:29,381 INFO org.apache.hadoop.mapred.TaskRunner: task_0133_r_66_1 done; removing files. 2007-08-21 17:02:31,272 INFO org.apache.hadoop.mapred.TaskTracker: task_0133_r_80_2 0.0% reduce copy 2007-08-21 17:02:32,275 INFO org.apache.hadoop.mapred.TaskTracker: task_0133_r_80_2 0.0% reduce copy 2007-08-21 17:02:33,277 INFO org.apache.hadoop.mapred.TaskTracker: task_0133_r_80_2 0.0% reduce copy ... [task_0133_r_80_2 continue to run] Of course the JobTracker kept on complaining 2007-08-22 19:06:37,880 INFO org.apache.hadoop.mapred.JobTracker: Serious problem. While updating status, cannot find taskid task_0133_r_80_2 2007-08-22 19:06:38,124 INFO org.apache.hadoop.mapred.JobTracker: Serious problem. While updating status, cannot find taskid task_0133_r_80_2 2007-08-22 19:06:47,885 INFO org.apache.hadoop.mapred.JobTracker: Serious problem. While updating status, cannot find taskid task_0133_r_80_2 -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-338) Need more complete API of JobClient class
[ https://issues.apache.org/jira/browse/MAPREDUCE-338?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-338. Resolution: Fixed I'm going to close this as stale. We're now 7 years on and many API changes later, including getting the information being asked for in this JIRA. Need more complete API of JobClient class - Key: MAPREDUCE-338 URL: https://issues.apache.org/jira/browse/MAPREDUCE-338 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Runping Qi We need a programmatic way to find out the information about a map/reduce cluster and the jobs on the cluster. The current API is not complete. In particular, the following API functions are needed: 1. jobs() currently, there is an API function JobsToComplete, which returns running/waiting jobs only. jobs() should return the complete list. 2. TaskReport[] getMap/ReduceTaskReports(String jobid) 3. getStartTime() 4. getJobStatus(String jobid); 5. getJobProfile(String jobid); -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-2841) Task level native optimization
[ https://issues.apache.org/jira/browse/MAPREDUCE-2841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065207#comment-14065207 ] Todd Lipcon commented on MAPREDUCE-2841: Thanks Sean. This patch looks better. I committed it as the initial import onto the new feature branch (MR-2841). I had some issues building on my Ubuntu 13.10 system, but one of the purposes of the feature branch is to be able to iterate on it more collaboratively. I'll file a couple of subtasks for the issues I'm running into on my box. Task level native optimization -- Key: MAPREDUCE-2841 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2841 Project: Hadoop Map/Reduce Issue Type: Improvement Components: task Environment: x86-64 Linux/Unix Reporter: Binglin Chang Assignee: Sean Zhong Attachments: DESIGN.html, MAPREDUCE-2841.v1.patch, MAPREDUCE-2841.v2.patch, dualpivot-0.patch, dualpivotv20-0.patch, fb-shuffle.patch, hadoop-3.0-mapreduce-2841-2014-7-17.patch I'm recently working on native optimization for MapTask based on JNI. The basic idea is that, add a NativeMapOutputCollector to handle k/v pairs emitted by mapper, therefore sort, spill, IFile serialization can all be done in native code, preliminary test(on Xeon E5410, jdk6u24) showed promising results: 1. Sort is about 3x-10x as fast as java(only binary string compare is supported) 2. IFile serialization speed is about 3x of java, about 500MB/s, if hardware CRC32C is used, things can get much faster(1G/ 3. Merge code is not completed yet, so the test use enough io.sort.mb to prevent mid-spill This leads to a total speed up of 2x~3x for the whole MapTask, if IdentityMapper(mapper does nothing) is used There are limitations of course, currently only Text and BytesWritable is supported, and I have not think through many things right now, such as how to support map side combine. I had some discussion with somebody familiar with hive, it seems that these limitations won't be much problem for Hive to benefit from those optimizations, at least. Advices or discussions about improving compatibility are most welcome:) Currently NativeMapOutputCollector has a static method called canEnable(), which checks if key/value type, comparator type, combiner are all compatible, then MapTask can choose to enable NativeMapOutputCollector. This is only a preliminary test, more work need to be done. I expect better final results, and I believe similar optimization can be adopt to reduce task and shuffle too. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-210) want InputFormat for zip files
[ https://issues.apache.org/jira/browse/MAPREDUCE-210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065216#comment-14065216 ] Allen Wittenauer commented on MAPREDUCE-210: Ping! want InputFormat for zip files -- Key: MAPREDUCE-210 URL: https://issues.apache.org/jira/browse/MAPREDUCE-210 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Doug Cutting Assignee: indrajit Attachments: ZipInputFormat_fixed.patch HDFS is inefficient with large numbers of small files. Thus one might pack many small files into large, compressed, archives. But, for efficient map-reduce operation, it is desireable to be able to split inputs into smaller chunks, with one or more small original file per split. The zip format, unlike tar, permits enumeration of files in the archive without scanning the entire archive. Thus a zip InputFormat could efficiently permit splitting large archives into splits that contain one or more archived files. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Created] (MAPREDUCE-5975) Fix native-task build on Ubuntu 13.10
Todd Lipcon created MAPREDUCE-5975: -- Summary: Fix native-task build on Ubuntu 13.10 Key: MAPREDUCE-5975 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5975 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Reporter: Todd Lipcon Assignee: Todd Lipcon Priority: Blocker I'm having some issues building the native-task branch on my Ubuntu 13.10 box. This JIRA is to figure out and fix whatever's going on. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Created] (MAPREDUCE-5976) native-task should not fail to build if snappy is missing
Todd Lipcon created MAPREDUCE-5976: -- Summary: native-task should not fail to build if snappy is missing Key: MAPREDUCE-5976 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5976 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Reporter: Todd Lipcon Assignee: Sean Zhong Other native parts of Hadoop will automatically disable snappy support if snappy is not present and -Drequire.snappy is not passed. native-task should do the same. (right now, it fails to build if snappy is missing) -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-428) Reducers reported completion % is generally incorrect when consuming compressed map outputs
[ https://issues.apache.org/jira/browse/MAPREDUCE-428?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-428. Resolution: Duplicate Reducers reported completion % is generally incorrect when consuming compressed map outputs --- Key: MAPREDUCE-428 URL: https://issues.apache.org/jira/browse/MAPREDUCE-428 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Riccardo Boscolo Priority: Minor When processing compressed map outputs, reducers often report over 100% completion (up to 220%). This is regardless of the compression codec and of whether native compression is used or not. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-209) Support for metrics aggregation module in JobTracker
[ https://issues.apache.org/jira/browse/MAPREDUCE-209?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-209. Resolution: Fixed Closing this as stale. Support for metrics aggregation module in JobTracker Key: MAPREDUCE-209 URL: https://issues.apache.org/jira/browse/MAPREDUCE-209 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Senthil Subramanian JobTracker should support starting up and shutting down a generic metrics aggregation module. We are currently thinking about plugging in a module that gets time series data from the task trackers, aggregates it and log this data into a global DFS so that it is analysed later (even after the map reduce cluster is shutdown). Some of this data can also be plotted on the JobTracker UI in realtime. This is particulary useful for analyzing data from dynamic mapreduce cluster like the ones deployed using HOD. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Created] (MAPREDUCE-5977) Fix or suppress native-task gcc warnings
Todd Lipcon created MAPREDUCE-5977: -- Summary: Fix or suppress native-task gcc warnings Key: MAPREDUCE-5977 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5977 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Reporter: Todd Lipcon Assignee: Todd Lipcon Currently, building the native task code on gcc 4.8 has a fair number of warnings. We should fix or suppress them so that new warnings are easier to see. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5975) Fix native-task build on Ubuntu 13.10
[ https://issues.apache.org/jira/browse/MAPREDUCE-5975?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Todd Lipcon updated MAPREDUCE-5975: --- Attachment: mr-5975.txt Trivial patch to add some missing unistd.h includes which were necessary to build on my box. Fix native-task build on Ubuntu 13.10 - Key: MAPREDUCE-5975 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5975 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Reporter: Todd Lipcon Assignee: Todd Lipcon Priority: Blocker Attachments: mr-5975.txt I'm having some issues building the native-task branch on my Ubuntu 13.10 box. This JIRA is to figure out and fix whatever's going on. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5975) Fix native-task build on Ubuntu 13.10
[ https://issues.apache.org/jira/browse/MAPREDUCE-5975?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065248#comment-14065248 ] Todd Lipcon commented on MAPREDUCE-5975: [~clockfly] and [~decster] -- since you guys are branch committers, you can review (and +1) patches targeted towards the native task feature branch. Mind taking a look at this one (and the others that I am filing?) Fix native-task build on Ubuntu 13.10 - Key: MAPREDUCE-5975 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5975 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Reporter: Todd Lipcon Assignee: Todd Lipcon Priority: Blocker Attachments: mr-5975.txt I'm having some issues building the native-task branch on my Ubuntu 13.10 box. This JIRA is to figure out and fix whatever's going on. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-2833) Job Tracker needs to collect more job/task execution stats and save them to DFS file
[ https://issues.apache.org/jira/browse/MAPREDUCE-2833?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-2833. - Resolution: Fixed I'm closing this as fixed since the history files pretty much over this request. Job Tracker needs to collect more job/task execution stats and save them to DFS file Key: MAPREDUCE-2833 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2833 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Runping Qi Labels: newbie In order to facilitate offline analysis on the dynamic behaviors and performance characterics of map/reduce jobs, we need the job tracker to collect some data about jobs and save them to DFS files. Some data are in time series form, and some are not. Below is a preliminary list of desired data. Some of them are already available in the current job trackers. Some are new. For each map/reduce job, we need the following non time series data: 1. jobid, jobname, number of mappers, number of reducers, start time, end time, end of mapper phase 2. Average (median, min, max) of successful mapper execution time, input/output records/bytes 3. Average (median, min, max) of uncessful mapper execution time, input/output records/bytes 4.Total mapper retries, max, average number of re-tries per mapper 5. The reasons for mapper task fails. 6. Average (median, min, max) of successful reducer execution time, input/output reocrds/bytes Execution time is the difference between the sort end time and the task end time 7. Average (median, min, max) of successful copy time (from the mapper phase end time to the sort start time). 8. Average (median, min, max) of successful sorting time for successful reducers 9. Average (median, min, max) of unsuccessful reducer execution time (from the end of mapper phase or the start of the task, whichever later, to the end of task) 10. Total reducer retries, max, average number of per reducer retries 11. The reasons for reducer task fails (user code error, lost tracker, failed to write to DFS, etc.) For each map/reduce job, we collect the following time series data (with one minute interval): 1. Numbers of pending mappers, reducers 2. Number of running mappers, reducers For the job tracker, we need the following data: 1. Number of trackers 2. Start time 3. End time 4. The list of map reduce jobs (their ids, starttime/endtime) The following time series data (with one minute interval): 1. The number of running jobs 2. The numbers of running mappers/reducers 3. The number pending mappers/reducers The data collection should be optional. That is, a job tracker can turn off such data collection, and in that case, it should not pay the cost. The job tracker should organize the in memory version of the collected data in such a way that: 1. it does not consume excessive amount of memory 2. the data may be suitable for presenting through the Web status pages. The data saved on DFS files should be in hadoop record format. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-230) Need to document the controls for sorting and grouping into the reduce
[ https://issues.apache.org/jira/browse/MAPREDUCE-230?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-230. Resolution: Won't Fix This ship sailed a looong time ago. Need to document the controls for sorting and grouping into the reduce -- Key: MAPREDUCE-230 URL: https://issues.apache.org/jira/browse/MAPREDUCE-230 Project: Hadoop Map/Reduce Issue Type: Task Reporter: Owen O'Malley Assignee: Arun C Murthy The JavaDoc for the Reducer should document how to control the sort order of keys and values via the JobConf methods: {code} setOutputKeyComparatorClass setOutputValueGroupingComparator {code} Both methods desperately need better names. (I'd vote for setKeySortingComparator and setKeyGroupingComparator.) -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-272) Job tracker should report the number of splits that are local to some task trackers
[ https://issues.apache.org/jira/browse/MAPREDUCE-272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065261#comment-14065261 ] Allen Wittenauer commented on MAPREDUCE-272: Don't we essentially have this information logged now? Or am I missing something here? Job tracker should report the number of splits that are local to some task trackers --- Key: MAPREDUCE-272 URL: https://issues.apache.org/jira/browse/MAPREDUCE-272 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Runping Qi Assignee: Runping Qi Attachments: hadoop-2015.txt Right now, jon tracker keeps track the number of launched mappers with local data. However, it is not clear how many mappers that are potentially be launched with data locality. This information is readily available in Job Tracker. It is just a matter to create a separate global counter and set it at the Job Tracker initialization time. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-129) job_null_0001 in jobid
[ https://issues.apache.org/jira/browse/MAPREDUCE-129?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-129. Resolution: Fixed job_null_0001 in jobid -- Key: MAPREDUCE-129 URL: https://issues.apache.org/jira/browse/MAPREDUCE-129 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Koji Noguchi Assignee: Owen O'Malley When I submit a job before jobtracker is fully up, I occasionally get jobid of job_null_0001. [knoguchi ]$ hadoop -jar ... 07/10/12 00:15:07 INFO mapred.FileInputFormat: Total input paths to process : 4 07/10/12 00:15:08 INFO mapred.JobClient: Running job: *job_null_0001* 07/10/12 00:15:09 INFO mapred.JobClient: map 0% reduce 0% -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5974) Allow map output collector fallback
[ https://issues.apache.org/jira/browse/MAPREDUCE-5974?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065268#comment-14065268 ] Hadoop QA commented on MAPREDUCE-5974: -- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12656299/mapreduce-5974.txt against trunk revision . {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:red}-1 tests included{color}. The patch doesn't appear to include any new or modified tests. Please justify why no new tests are needed for this patch. Also please list what manual steps were performed to verify this patch. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:green}+1 javadoc{color}. There were no new javadoc warning messages. {color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse. {color:green}+1 findbugs{color}. The patch does not introduce any new Findbugs (version 2.0.3) warnings. {color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings. {color:green}+1 core tests{color}. The patch passed unit tests in hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core. {color:green}+1 contrib tests{color}. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4749//testReport/ Console output: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4749//console This message is automatically generated. Allow map output collector fallback --- Key: MAPREDUCE-5974 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5974 Project: Hadoop Map/Reduce Issue Type: Sub-task Components: task Affects Versions: 2.6.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Attachments: mapreduce-5974.txt Currently we only allow specifying a single MapOutputCollector implementation class in a job. It would be nice to allow a comma-separated list of classes: we should try each collector implementation in the user-specified order until we find one that can be successfully instantiated and initted. This is useful for cases where a particular optimized collector implementation cannot operate on all key/value types, or requires native code. The cluster administrator can configure the cluster to try to use the optimized collector and fall back to the default collector. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-492) Pending, running, completed tasks should also be shown as percentage
[ https://issues.apache.org/jira/browse/MAPREDUCE-492?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-492. Resolution: Fixed A different percentage was committed eons ago. Pending, running, completed tasks should also be shown as percentage Key: MAPREDUCE-492 URL: https://issues.apache.org/jira/browse/MAPREDUCE-492 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Amar Kamat Assignee: Amar Kamat Priority: Minor Attachments: HADOOP-2099.patch, percent.png -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Assigned] (MAPREDUCE-5962) Support CRC32C in IFile
[ https://issues.apache.org/jira/browse/MAPREDUCE-5962?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Todd Lipcon reassigned MAPREDUCE-5962: -- Assignee: Todd Lipcon (was: James Thomas) Support CRC32C in IFile --- Key: MAPREDUCE-5962 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5962 Project: Hadoop Map/Reduce Issue Type: Improvement Components: performance, task Affects Versions: 2.5.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Currently, the IFile format used by the MR shuffle checksums all data using the zlib CRC32 polynomial. If we allow use of CRC32C instead, we can get a large reduction in CPU usage by leveraging the native hardware CRC32C implementation (approx half a second of CPU time savings per GB checksummed). -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-2830) Document config parameters for each Map-Reduce class/interface
[ https://issues.apache.org/jira/browse/MAPREDUCE-2830?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065280#comment-14065280 ] Allen Wittenauer commented on MAPREDUCE-2830: - Ping! Document config parameters for each Map-Reduce class/interface -- Key: MAPREDUCE-2830 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2830 Project: Hadoop Map/Reduce Issue Type: Improvement Components: documentation Reporter: Arun C Murthy Labels: newbie I propose we add a table in the javadoc for each user-facing Map-Reduce interface/class which lists, and provides details, of each and every config parameter which has any bearing on that interface/class. Clearly some parameters affect more than one place and they should be put in more than one table. For e.g. Mapper - io.sort.mb, io.sort.factor Reducer - fs.inmemory.size.mb ... etc. It would very nice to explain how it interacts with the framework and rest of config params etc. Thoughts? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5962) Support CRC32C in IFile
[ https://issues.apache.org/jira/browse/MAPREDUCE-5962?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Todd Lipcon updated MAPREDUCE-5962: --- Attachment: mapreduce-5962.txt Attached patch adds a new configuration to set the IFile checksum type. I changed the default to CRC32C since it's much faster if you have the native libraries available. I don't believe this is an incompatible change, since IFiles are only used internal to a single job (written by map, read by reduce). So, one would never have a different version reader compared to writer. That said, if anyone has any issues with this, they can configure the default back to CRC32 cluster-wide. Support CRC32C in IFile --- Key: MAPREDUCE-5962 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5962 Project: Hadoop Map/Reduce Issue Type: Improvement Components: performance, task Affects Versions: 2.5.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Attachments: mapreduce-5962.txt Currently, the IFile format used by the MR shuffle checksums all data using the zlib CRC32 polynomial. If we allow use of CRC32C instead, we can get a large reduction in CPU usage by leveraging the native hardware CRC32C implementation (approx half a second of CPU time savings per GB checksummed). -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065293#comment-14065293 ] Hadoop QA commented on MAPREDUCE-5910: -- {color:green}+1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12656302/MAPREDUCE-5910.4.patch against trunk revision . {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 1 new or modified test files. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:green}+1 javadoc{color}. There were no new javadoc warning messages. {color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse. {color:green}+1 findbugs{color}. The patch does not introduce any new Findbugs (version 2.0.3) warnings. {color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings. {color:green}+1 core tests{color}. The patch passed unit tests in hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app. {color:green}+1 contrib tests{color}. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4748//testReport/ Console output: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/4748//console This message is automatically generated. MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch, MAPREDUCE-5910.4.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-5962) Support CRC32C in IFile
[ https://issues.apache.org/jira/browse/MAPREDUCE-5962?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Todd Lipcon updated MAPREDUCE-5962: --- Status: Patch Available (was: Open) Support CRC32C in IFile --- Key: MAPREDUCE-5962 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5962 Project: Hadoop Map/Reduce Issue Type: Improvement Components: performance, task Affects Versions: 2.5.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Attachments: mapreduce-5962.txt Currently, the IFile format used by the MR shuffle checksums all data using the zlib CRC32 polynomial. If we allow use of CRC32C instead, we can get a large reduction in CPU usage by leveraging the native hardware CRC32C implementation (approx half a second of CPU time savings per GB checksummed). -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5962) Support CRC32C in IFile
[ https://issues.apache.org/jira/browse/MAPREDUCE-5962?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065291#comment-14065291 ] Todd Lipcon commented on MAPREDUCE-5962: (fwiw this depends on [~james.thomas]'s work to enable native checksumming on byte arrays. So we won't see an immediate benefit, but will once that patch is done) Support CRC32C in IFile --- Key: MAPREDUCE-5962 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5962 Project: Hadoop Map/Reduce Issue Type: Improvement Components: performance, task Affects Versions: 2.5.0 Reporter: Todd Lipcon Assignee: Todd Lipcon Attachments: mapreduce-5962.txt Currently, the IFile format used by the MR shuffle checksums all data using the zlib CRC32 polynomial. If we allow use of CRC32C instead, we can get a large reduction in CPU usage by leveraging the native hardware CRC32C implementation (approx half a second of CPU time savings per GB checksummed). -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-108) Blacklisted hosts may not be able to serve map outputs
[ https://issues.apache.org/jira/browse/MAPREDUCE-108?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-108. Resolution: Incomplete I'm going to close this out as stale. Blacklisting has undergone quite a few changes in the past 6-7 years. Blacklisted hosts may not be able to serve map outputs -- Key: MAPREDUCE-108 URL: https://issues.apache.org/jira/browse/MAPREDUCE-108 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Runping Qi Assignee: Amar Kamat Attachments: HADOOP-2175-v1.1.patch, HADOOP-2175-v1.patch, HADOOP-2175-v2.patch, HADOOP-2175-v2.patch After a node fails 4 mappers (tasks), it is added to blacklist thus it will no longer accept tasks. But, it will continue serve the map outputs of any mappers that ran successfully there. However, the node may not be able serve the map outputs either. This will cause the reducers to mark the corresponding map outputs as from slow hosts, but continue to try to get the map outputs from that node. This may lead to waiting forever. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-155) Each task tracker should not execute more than one speculative task
[ https://issues.apache.org/jira/browse/MAPREDUCE-155?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-155. Resolution: Incomplete Closing this as stale since spec exec has undergone quite a bit of overhauling which I'm confident included fixing this... at least, I've never seen a job get multiple versions of the same task assigned to the same node in practice. Each task tracker should not execute more than one speculative task --- Key: MAPREDUCE-155 URL: https://issues.apache.org/jira/browse/MAPREDUCE-155 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Runping Qi I noticed that sometimes, a tasktracker started 2 or three speculative mapper tasks. That seems counter productive. You want to speculative execution complete as soon as possible. Thus, it is better to spread speculative execution over multiple trackers. A simple way to achieve that is to limit the number of speculative eecution concurrently. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-636) Inconsistency in config parameter for RandomWriter
[ https://issues.apache.org/jira/browse/MAPREDUCE-636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer updated MAPREDUCE-636: --- Labels: newbie (was: ) Inconsistency in config parameter for RandomWriter -- Key: MAPREDUCE-636 URL: https://issues.apache.org/jira/browse/MAPREDUCE-636 Project: Hadoop Map/Reduce Issue Type: Bug Components: examples Reporter: Arun C Murthy Assignee: Arun C Murthy Priority: Minor Labels: newbie All configuration parameters for RandomWriter start with the suffix {{test.randomwrite}}, except for {{test.randomwriter.maps_per_host}}. Minor inconsistency to fix. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-1769) Streaming command should be able to take its output to a file, rather then to stdout
[ https://issues.apache.org/jira/browse/MAPREDUCE-1769?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-1769. - Resolution: Won't Fix Same reasoning as its buddy jira. Streaming command should be able to take its output to a file, rather then to stdout -- Key: MAPREDUCE-1769 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1769 Project: Hadoop Map/Reduce Issue Type: Improvement Components: contrib/streaming Reporter: arkady borkovsky In some cases, especially when a streaming command is a 3rd party or legacy application, it is impossible of inconvenient to make it write its output to stdout The command may require that the ouput file name is specified as a command line option, or the output file name is hard coded. Related to https://issues.apache.org/jira/browse/HADOOP-2235 -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065309#comment-14065309 ] Jian He commented on MAPREDUCE-5910: committing this. MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch, MAPREDUCE-5910.4.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-1767) Steaming infrastructures should provide statisics about job
[ https://issues.apache.org/jira/browse/MAPREDUCE-1767?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-1767. - Resolution: Fixed All of this information is available via other means. Steaming infrastructures should provide statisics about job --- Key: MAPREDUCE-1767 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1767 Project: Hadoop Map/Reduce Issue Type: Improvement Components: contrib/streaming Reporter: arkady borkovsky This should include -- the commands (mapper and reducer commands) executed -- time information (e.g. min, max, and avg start time, end time, elapsed time for tasks, total elapsed time ) -- sizes -- bytes and records, min, max, avg per task and total, input and output -- information about input and output data sets (all output data sets, if there are several) -- all user counters (when they are implemented for streaming) the information should be stored in a file -- e.g. in the working directory from where the job was launched, with a name derived from the job name -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-611) Streaming infrastructure should report information about runtime errors
[ https://issues.apache.org/jira/browse/MAPREDUCE-611?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-611. Resolution: Won't Fix The stderr log was built to handle this type of thing. Closing as won't fix. Streaming infrastructure should report information about runtime errors Key: MAPREDUCE-611 URL: https://issues.apache.org/jira/browse/MAPREDUCE-611 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: arkady borkovsky For example, if the streaming command is Perl script an syntax error or a runtime error occurs during script execution, the error message (the stack trace) should be reported to the user, separate from and in addition to the rest of the logs and the stderr output. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-2829) TestMiniMRMapRedDebugScript times out
[ https://issues.apache.org/jira/browse/MAPREDUCE-2829?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-2829. - Resolution: Fixed Yup. Definitely. Closing. TestMiniMRMapRedDebugScript times out - Key: MAPREDUCE-2829 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2829 Project: Hadoop Map/Reduce Issue Type: Bug Environment: Linux Reporter: Konstantin Shvachko Attachments: Hadoop-2260.log, testrun-2260.log I am running TestMiniMRMapRedDebugScript from trunc. This is what I see in the stdout: {code} 2007-11-22 02:21:23,494 WARN conf.Configuration (Configuration.java:loadResource(808)) - hadoop/build/test/mapred/local/1_0/taskTracker/jobcache/job_200711220217_0001/task_200711220217_0001_m_00_0/job.xml:a attempt to override final parameter: hadoop.tmp.dir; Ignoring. 2007-11-22 02:21:28,940 INFO jvm.JvmMetrics (JvmMetrics.java:init(56)) - Initializing JVM Metrics with processName=MAP, sessionId= 2007-11-22 02:22:09,504 INFO mapred.MapTask (MapTask.java:run(127)) - numReduceTasks: 0 2007-11-22 02:22:42,434 WARN mapred.TaskTracker (TaskTracker.java:main(1982)) - Error running child java.io.IOException at org.apache.hadoop.mapred.TestMiniMRMapRedDebugScript$MapClass.map(TestMiniMRMapRedDebugScript.java:41) at org.apache.hadoop.mapred.TestMiniMRMapRedDebugScript$MapClass.map(TestMiniMRMapRedDebugScript.java:35) at org.apache.hadoop.mapred.MapRunner.run(MapRunner.java:50) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:192) at org.apache.hadoop.mapred.TaskTracker$Child.main(TaskTracker.java:1977) {code} Stderr and debugout both say: Bailing out. BTW on Windows everything works just fine. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-229) Provide a command line option to check if a Hadoop jobtracker is idle
[ https://issues.apache.org/jira/browse/MAPREDUCE-229?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-229. Resolution: Fixed I'm going to close this as fixed. Clients can ask the JT if they are currently busy. By doing this periodically, they can build an idle time. Provide a command line option to check if a Hadoop jobtracker is idle - Key: MAPREDUCE-229 URL: https://issues.apache.org/jira/browse/MAPREDUCE-229 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Hemanth Yamijala This is an RFE for providing a way to determine from the hadoop command line whether a jobtracker is idle. One possibility is to have something like hadoop jobtracker -idle time. Hadoop would return true (maybe via some stdout output) if the jobtracker had no work to do (jobs running / prepared) since time seconds, false otherwise. This would be useful for management / provisioning systems like Hadoop-On-Demand [HADOOP-1301], which can then deallocate the idle, provisioned clusters automatically, and release resources. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-620) Streaming: support local execution
[ https://issues.apache.org/jira/browse/MAPREDUCE-620?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-620. Resolution: Won't Fix Unix already provides this functionality. Streaming: support local execution -- Key: MAPREDUCE-620 URL: https://issues.apache.org/jira/browse/MAPREDUCE-620 Project: Hadoop Map/Reduce Issue Type: Improvement Components: contrib/streaming Reporter: arkady borkovsky For streaming, local execution does not involve hadoop. It is just hdfs -cat input | mapper-command | sort | reducer command While a user can do this herself, having an option to do this by using the infrastructure would greatly simplify user script and and make it easier to ensure that the process will run on the cluster as expected. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-598) Streaming: better conrol over input splits
[ https://issues.apache.org/jira/browse/MAPREDUCE-598?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-598. Resolution: Fixed This has been fixed in a variety of use cases. Closing. Streaming: better conrol over input splits -- Key: MAPREDUCE-598 URL: https://issues.apache.org/jira/browse/MAPREDUCE-598 Project: Hadoop Map/Reduce Issue Type: Improvement Components: contrib/streaming Reporter: arkady borkovsky In steaming, the map command usually expect to receive it's input uninterpreted -- just as it is stored in DFS. However, the split (the beginning and the end of the portion of data that goes to a single map task) is often important and is not any line break. Often the input consists of multi-line docments -- e.g. in XML. There should be a way to specify a pattern that separates logical records. Existing Streaming XML record reader kind of provides this functionality. However, it is accepted that Streaming XML is a hack and needs to be replaced -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-5910) MRAppMaster should handle Resync from RM instead of shutting down.
[ https://issues.apache.org/jira/browse/MAPREDUCE-5910?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065361#comment-14065361 ] Hudson commented on MAPREDUCE-5910: --- FAILURE: Integrated in Hadoop-trunk-Commit #5902 (See [https://builds.apache.org/job/Hadoop-trunk-Commit/5902/]) MAPREDUCE-5910. Make MR AM resync with RM in case of work-preserving RM-restart. Contributed by Rohith (jianhe: http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1611434) * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/local/LocalContainerAllocator.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/rm/RMCommunicator.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/rm/RMContainerAllocator.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/rm/RMContainerRequestor.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapreduce/v2/app/rm/TestRMContainerAllocator.java MRAppMaster should handle Resync from RM instead of shutting down. -- Key: MAPREDUCE-5910 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5910 Project: Hadoop Map/Reduce Issue Type: Task Components: applicationmaster Reporter: Rohith Assignee: Rohith Attachments: MAPREDUCE-5910.1.patch, MAPREDUCE-5910.2.patch, MAPREDUCE-5910.3.patch, MAPREDUCE-5910.4.patch The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The MRAppMaster behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-618) Need to be able to re-run specific map tasks (when -reducer NONE)
[ https://issues.apache.org/jira/browse/MAPREDUCE-618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065373#comment-14065373 ] Allen Wittenauer commented on MAPREDUCE-618: I'm tempted to close this because retry logic is way better now. However... isn't this essentially the same request as the various (real) preemption jiras? Need to be able to re-run specific map tasks (when -reducer NONE) - Key: MAPREDUCE-618 URL: https://issues.apache.org/jira/browse/MAPREDUCE-618 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: arkady borkovsky Sometimes, a few map tasks fail and -reducer NONE. It should be possible to rerun the failed map tasks . There are several failure modes: * a task is hanging, so the job is killed * from the infrastructure perspective, the task has completed successfully , but it failed to produces correct result * failed in the proper Hadoop sense It is often too expensive to rerun the whole job. And for larger jobs, chances are each run will have a few failed tasks. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-386) The combiner in pipes is closed before the last values are passed in.
[ https://issues.apache.org/jira/browse/MAPREDUCE-386?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-386. Resolution: Fixed Stale issue. I'm sure this has been fixed by now! The combiner in pipes is closed before the last values are passed in. - Key: MAPREDUCE-386 URL: https://issues.apache.org/jira/browse/MAPREDUCE-386 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Owen O'Malley Assignee: Owen O'Malley Currently the last spill is sent to the combiner after the close method is called. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-594) Streaming: org.apache.hadoop.mapred.lib.IdentityMapper should not inserted unnecessary keys
[ https://issues.apache.org/jira/browse/MAPREDUCE-594?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-594. Resolution: Fixed Closing this as fixed for a variety of reasons (alternative provided, you can now provide your own input format, etc, etc) Streaming: org.apache.hadoop.mapred.lib.IdentityMapper should not inserted unnecessary keys --- Key: MAPREDUCE-594 URL: https://issues.apache.org/jira/browse/MAPREDUCE-594 Project: Hadoop Map/Reduce Issue Type: Bug Components: contrib/streaming Reporter: arkady borkovsky When streaming command specifies -mapper org.apache.hadoop.mapred.lib.IdentityMapper the reducer should receive exactly the same text lines as where present in the input. The only modification is the reordering the input. Currently, org.apache.hadoop.mapred.lib.IdentityMapper inserts ofsets in the input as keys. Which renders it useless. Moreover, in the latest release org.apache.hadoop.mapred.lib.IdentityMapper just crashes: java.io.IOException: Type mismatch in key from map: e xpected org.apache.hadoop.io.Text, recieved org.apache.hadoop.io.LongWritable at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.collect(MapTask.java:331) at org.apache.hadoop.mapred.lib.IdentityMapper.map(IdentityMapper.java:40) at org.apache.hadoop.mapred.MapRunner.run(MapRunner.java:50) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:192) at org.apache.hadoop.mapred.TaskTracker$Child.main(TaskTracker.java:1760) (I open only one bug, as it is broken anyway, the new behavior does not actually make it any worse than before) -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-455) Hadoop needs a better XML Input
[ https://issues.apache.org/jira/browse/MAPREDUCE-455?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065392#comment-14065392 ] Allen Wittenauer commented on MAPREDUCE-455: Is it time to revisit this? Hadoop needs a better XML Input --- Key: MAPREDUCE-455 URL: https://issues.apache.org/jira/browse/MAPREDUCE-455 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Alan Ho Priority: Minor Attachments: HADOOP-2439Patch.patch Hadoop does not have a good XML parser for XML input. The XML parser in the streaming class is fairly difficult to work with and doesn't have proper test cases around it. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-215) Improve facilities for job-control, job-queues etc.
[ https://issues.apache.org/jira/browse/MAPREDUCE-215?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-215. Resolution: Fixed I'm going to close as fixed for a variety of reasons: - some of these features are now native to Hadoop - some of these features are now part of Oozie, Azkaban, etc - some of these features are part of Tez Improve facilities for job-control, job-queues etc. --- Key: MAPREDUCE-215 URL: https://issues.apache.org/jira/browse/MAPREDUCE-215 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Arun C Murthy Today, Map-Reduce has _some_ support for job-control - basically JobClient provides a facility to monitor jobs, one can setup a job-ending notification and there is {{JobControl}}. Links: http://lucene.apache.org/hadoop/docs/r0.15.1/mapred_tutorial.html#Job+Control http://lucene.apache.org/hadoop/docs/r0.15.1/mapred_tutorial.html#JobControl Looks like users could do more with better facilities for job-control and maybe more advanced features like job-queues etc. Lets discuss... -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-265) check permissions for job inputs and outputs
[ https://issues.apache.org/jira/browse/MAPREDUCE-265?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-265. Resolution: Fixed I believe this has been fixed. check permissions for job inputs and outputs Key: MAPREDUCE-265 URL: https://issues.apache.org/jira/browse/MAPREDUCE-265 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Doug Cutting Attachments: HADOOP-2528-0.patch, HADOOP-2528-1.patch On job submission, filesystem permissions should be checked to ensure that the input directory is readable and that the output directory is writable. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-186) TaskLogServlet returns 410 when trying to access log early in task life
[ https://issues.apache.org/jira/browse/MAPREDUCE-186?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-186. Resolution: Incomplete I'm going to close this out as stale. TaskLogServlet returns 410 when trying to access log early in task life --- Key: MAPREDUCE-186 URL: https://issues.apache.org/jira/browse/MAPREDUCE-186 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Michael Bieniosek Attachments: hadoop-2572.patch Early in a map task life, or for tasks that died quickly, the file $task/syslog might not exist. In this case, the TaskLogServlet gives a status 410. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-224) limit running tasks per job
[ https://issues.apache.org/jira/browse/MAPREDUCE-224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-224. Resolution: Duplicate I'm going to close this out as a duplicate of MAPREDEUCE-5583. limit running tasks per job --- Key: MAPREDUCE-224 URL: https://issues.apache.org/jira/browse/MAPREDUCE-224 Project: Hadoop Map/Reduce Issue Type: New Feature Reporter: Doug Cutting It should be possible to specify a limit to the number of tasks per job permitted to run simultaneously. If, for example, you have a cluster of 50 nodes, with 100 map task slots and 100 reduce task slots, and the configured limit is 25 simultaneous tasks/job, then four or more jobs will be able to run at a time. This will permit short jobs to pass longer-running jobs. This also avoids some problems we've seen with HOD, where nodes are underutilized in their tail, and it should permit improved input locality. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-574) Fix -file option in Streaming to use Distributed Cache
[ https://issues.apache.org/jira/browse/MAPREDUCE-574?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-574. Resolution: Fixed Stale. Fix -file option in Streaming to use Distributed Cache -- Key: MAPREDUCE-574 URL: https://issues.apache.org/jira/browse/MAPREDUCE-574 Project: Hadoop Map/Reduce Issue Type: Bug Components: distributed-cache Reporter: Amareshwari Sriramadasu Attachments: patch-2622.txt The -file option works by putting the script into the job's jar file by unjar-ing, copying and then jar-ing it again. We should rework the -file option to use the DistributedCache and the symlink option it provides. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-165) the map task output servlet doesn't protect against .. attacks
[ https://issues.apache.org/jira/browse/MAPREDUCE-165?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065441#comment-14065441 ] Allen Wittenauer commented on MAPREDUCE-165: I suspect this issue can be closed, but it would be good to have some verification. the map task output servlet doesn't protect against .. attacks Key: MAPREDUCE-165 URL: https://issues.apache.org/jira/browse/MAPREDUCE-165 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Owen O'Malley Labels: security The servlet we use to export the map outputs doesn't protect itself against .. attacks. However, because the code adds a /file.out.index and /file.out to it, it can only be used to read files with those names. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-165) the map task output servlet doesn't protect against .. attacks
[ https://issues.apache.org/jira/browse/MAPREDUCE-165?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer updated MAPREDUCE-165: --- Labels: security (was: ) the map task output servlet doesn't protect against .. attacks Key: MAPREDUCE-165 URL: https://issues.apache.org/jira/browse/MAPREDUCE-165 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Owen O'Malley Labels: security The servlet we use to export the map outputs doesn't protect itself against .. attacks. However, because the code adds a /file.out.index and /file.out to it, it can only be used to read files with those names. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-302) Maintaining cluster information across multiple job submissions
[ https://issues.apache.org/jira/browse/MAPREDUCE-302?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065445#comment-14065445 ] Allen Wittenauer commented on MAPREDUCE-302: I think YARN sort of makes this JIRA obsolete, but I'd like some verification before closing it. Maintaining cluster information across multiple job submissions --- Key: MAPREDUCE-302 URL: https://issues.apache.org/jira/browse/MAPREDUCE-302 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Lohit Vijayarenu Assignee: dhruba borthakur Could we have a way to maintain cluster state across multiple job submissions. Consider a scenario where we run multiple jobs in iteration on a cluster back to back. The nature of the job is same, but input/output might differ. Now, if a node is blacklisted in one iteration of job run, it would be useful to maintain this information and blacklist this node for next iteration of job as well. Another situation which we saw is, if there are failures less than mapred.map.max.attempts in each iterations few nodes are never marked for blacklisting. But in we consider two or three iterations, these nodes fail all jobs and should be taken out of cluster. This hampers overall performance of the job. Could have have config variables something which matches a job type (provided by user) and maintains the cluster status for that job type alone? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-165) the map task output servlet doesn't protect against .. attacks
[ https://issues.apache.org/jira/browse/MAPREDUCE-165?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer updated MAPREDUCE-165: --- Labels: newbie security (was: security) the map task output servlet doesn't protect against .. attacks Key: MAPREDUCE-165 URL: https://issues.apache.org/jira/browse/MAPREDUCE-165 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Owen O'Malley Labels: newbie, security The servlet we use to export the map outputs doesn't protect itself against .. attacks. However, because the code adds a /file.out.index and /file.out to it, it can only be used to read files with those names. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (MAPREDUCE-489) Force the task tracker to exit when the task is complete, prevents nodes from dying due to resource starvation from impropertly written map/reduce tasks
[ https://issues.apache.org/jira/browse/MAPREDUCE-489?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer updated MAPREDUCE-489: --- Labels: newbie (was: ) Force the task tracker to exit when the task is complete, prevents nodes from dying due to resource starvation from impropertly written map/reduce tasks Key: MAPREDUCE-489 URL: https://issues.apache.org/jira/browse/MAPREDUCE-489 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Jason Priority: Minor Labels: newbie Original Estimate: 1h Remaining Estimate: 1h We have map/reduce jobs that sometimes run additional threads that are not at daemon priority, and these threads prevent the Task from properly exiting. When enough of these accumulate, the node falls over. The included patch forces the Tasks to exit when completed. Index: src/java/org/apache/hadoop/mapred/TaskTracker.java === --- src/java/org/apache/hadoop/mapred/TaskTracker.java (revision 608611) +++ src/java/org/apache/hadoop/mapred/TaskTracker.java (working copy) @@ -1801,6 +1801,8 @@ // This assumes that on return from Task.run() // there is no more logging done. LogManager.shutdown(); + +System.exit(0); } } } -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-489) Force the task tracker to exit when the task is complete, prevents nodes from dying due to resource starvation from impropertly written map/reduce tasks
[ https://issues.apache.org/jira/browse/MAPREDUCE-489?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065462#comment-14065462 ] Allen Wittenauer commented on MAPREDUCE-489: Doing a quick pass through the source seems to indicate that we don't always call system.exit on the way out. We probably should. Force the task tracker to exit when the task is complete, prevents nodes from dying due to resource starvation from impropertly written map/reduce tasks Key: MAPREDUCE-489 URL: https://issues.apache.org/jira/browse/MAPREDUCE-489 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Jason Priority: Minor Labels: newbie Original Estimate: 1h Remaining Estimate: 1h We have map/reduce jobs that sometimes run additional threads that are not at daemon priority, and these threads prevent the Task from properly exiting. When enough of these accumulate, the node falls over. The included patch forces the Tasks to exit when completed. Index: src/java/org/apache/hadoop/mapred/TaskTracker.java === --- src/java/org/apache/hadoop/mapred/TaskTracker.java (revision 608611) +++ src/java/org/apache/hadoop/mapred/TaskTracker.java (working copy) @@ -1801,6 +1801,8 @@ // This assumes that on return from Task.run() // there is no more logging done. LogManager.shutdown(); + +System.exit(0); } } } -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-344) Remove dead code block in JobInProgress.completedTask
[ https://issues.apache.org/jira/browse/MAPREDUCE-344?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-344. Resolution: Fixed Stale. Remove dead code block in JobInProgress.completedTask -- Key: MAPREDUCE-344 URL: https://issues.apache.org/jira/browse/MAPREDUCE-344 Project: Hadoop Map/Reduce Issue Type: Improvement Reporter: Arun C Murthy Since the taskCommitThread ensures that one and only one task of a given TIP is marked as SUCCEEDED, we don't need the code block in JobInProgress.completedTask which checks if the TIP is complete and then just marks the task as complete: {noformat} // Sanity check: is the TIP already complete? if (tip.isComplete()) { // Mark this task as KILLED tip.alreadyCompletedTask(taskid); // Let the JobTracker cleanup this taskid if the job isn't running if (this.status.getRunState() != JobStatus.RUNNING) { jobtracker.markCompletedTaskAttempt(status.getTaskTracker(), taskid); } return false; } {noformat} -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Resolved] (MAPREDUCE-390) Corner case exists in detecting Java process deaths that might lead to orphan pipes processes lying around in memory
[ https://issues.apache.org/jira/browse/MAPREDUCE-390?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Allen Wittenauer resolved MAPREDUCE-390. Resolution: Fixed I believe this one has actually been resolved with the rewrite of the pipes interface. Corner case exists in detecting Java process deaths that might lead to orphan pipes processes lying around in memory Key: MAPREDUCE-390 URL: https://issues.apache.org/jira/browse/MAPREDUCE-390 Project: Hadoop Map/Reduce Issue Type: Bug Reporter: Devaraj Das Priority: Minor In HADOOP-2092, the child pipes process periodically pings the parent Java process to find out whether it is alive. The ping cycle is 5 seconds. Consider the following scenario: 1) The Java task dies at the beginning of the ping cycle 2) A new Java task starts and binds to the same port as the earlier Java task's port 3) The pipes process wakes up and does a ping - it will still be successful since the port number hasn't changed This will lead to orphan processes lying around in memory. The detection of parent process deaths can be made more reliable at least on Unix'ish platforms by checking whether the parent process ID is 1, and if so exit. This will take care of the most common platform that hadoop is run on. For non-unix platforms, the existing ping mechanism can be retained. Thoughts? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (MAPREDUCE-2593) Random read benchmark for DFS
[ https://issues.apache.org/jira/browse/MAPREDUCE-2593?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14065470#comment-14065470 ] Allen Wittenauer commented on MAPREDUCE-2593: - Ping! Random read benchmark for DFS - Key: MAPREDUCE-2593 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2593 Project: Hadoop Map/Reduce Issue Type: New Feature Components: mrv2 Reporter: Raghu Angadi Assignee: Dave Thompson Attachments: HDFS-236.patch, RndRead-TestDFSIO-061011.patch, RndRead-TestDFSIO-MR2593-trunk121211.patch, RndRead-TestDFSIO.patch We should have at least one random read benchmark that can be run with rest of Hadoop benchmarks regularly. Please provide benchmark ideas or requirements. -- This message was sent by Atlassian JIRA (v6.2#6252)