Hadoop之——错误及解决方案总结

转载请注明出处：http://blog.csdn.net/l1028386804/article/details/51576518

1，错误一：java.io.IOException: Incompatible clusterIDs时常出现在namenode重新格式化之后9 `7 k# I: L2 |9 U* @6 d

2014-04-29 14:32:53,877 FATAL org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed forblock pool Block pool BP-1480406410-192.168.1.181-1398701121586 (storage idDS-167510828-192.168.1.191-50010-1398750515421) service tohadoop-master/192.168.1.181:9000" J' |7 h2q( T& @$ h" s' B
java.io.IOException: Incompatible clusterIDs in/data/dfs/data: namenode clusterID = CID-d1448b9e-da0f-499e-b1d4-78cb18ecdebb;datanode clusterID = CID-ff0faa40-2940-4838-b321-98272eb0dee3! U8 t) L- F( @0 ~' H0 N9 I
        atorg.apache.hadoop.hdfs.server.datanode.DataStorage.doTransition(DataStorage.java:391)5 ~" a4 j4 o6 M7 ~* r
        atorg.apache.hadoop.hdfs.server.datanode.DataStorage.recoverTransitionRead(DataStorage.java:191)8 {* e. t; f7 ?# I8 I: \- v
        atorg.apache.hadoop.hdfs.server.datanode.DataStorage.recoverTransitionRead(DataStorage.java:219)
        atorg.apache.hadoop.hdfs.server.datanode.DataNode.initStorage(DataNode.java:837)
        atorg.apache.hadoop.hdfs.server.datanode.DataNode.initBlockPool(DataNode.java:808)9 l  e( o1 o  u# D
        atorg.apache.hadoop.hdfs.server.datanode.BPOfferService.verifyAndSetNamespaceInfo(BPOfferService.java:280)
        atorg.apache.hadoop.hdfs.server.datanode.BPServiceActor.connectToNNAndHandshake(BPServiceActor.java:222)
        atorg.apache.hadoop.hdfs.server.datanode.BPServiceActor.run(BPServiceActor.java:664)* j) }9 t/ x* ~
        atjava.lang.Thread.run(Thread.java:722)
2014-04-29 14:32:53,885 WARNorg.apache.hadoop.hdfs.server.datanode.DataNode: Ending block pool service for:Block pool BP-1480406410-192.168.1.181-1398701121586 (storage idDS-167510828-192.168.1.191-50010-1398750515421) service tohadoop-master/192.168.1.181:90002 V9 G- G3 f* L
2014-04-29 14:32:53,889 INFOorg.apache.hadoop.hdfs.server.datanode.DataNode: Removed Block pool BP-1480406410-192.168.1.181-1398701121586(storage id DS-167510828-192.168.1.191-50010-1398750515421)
2014-04-29 14:32:55,897 WARNorg.apache.hadoop.hdfs.server.datanode.DataNode: Exiting Datanode

原因： 每次 namenode format 会重新创建一个 namenodeId, 而 data 目录包含了上次 format 时的 id,namenode format 清空了 namenode 下的数据 , 但是没有清空 datanode 下的数据 , 导致启动时失败 , 所要做的就是每次 fotmat 前 , 清空 data 下的所有目录 .
: d6 E2 t& M" g7 a* q3 l, H

解决办法：停掉集群，删除问题节点的data目录下的所有内容。即hdfs-site.xml文件中配置的dfs.data.dir目录。重新格式化namenode。

另一个更省事的办法：先停掉集群，然后将datanode节点目录/dfs/data/current/VERSION中的修改为与namenode一致即可。

2，错误：org.apache.hadoop.yarn.exceptions.YarnException:Unauthorized request to start container
. Y& }; L% ^' M8 H6 I5 T
14/04/29 02:45:07 INFO mapreduce.Job: Jobjob_1398704073313_0021 failed with state FAILED due to: Applicationapplication_1398704073313_0021 failed 2 times due to Error launchingappattempt_1398704073313_0021_000002. Got exception:org.apache.hadoop.yarn.exceptions.YarnException: Unauthorized request to startcontainer. ' F4 }0 C* `/ y# L9 A
This token is expired. current time is1398762692768 found 1398711306590
        atsun.reflect.GeneratedConstructorAccessor30.newInstance(Unknown Source)5 [& j' H( j0 j9 _4 ?7 o; q6 m
        at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)7 g- {( w6 [( N) s" `
        atjava.lang.reflect.Constructor.newInstance(Constructor.java:525)
        atorg.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBImpl.java:152)& n/ J* ]3 |& a2 q) t* g
        atorg.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:106), B4 u7 G  f, f  d6 H
        atorg.apache.hadoop.yarn.server.resourcemanager.amlauncher.AMLauncher.launch(AMLauncher.java:122)0 o+ [. u$ O; S' S- ?3 t. y
        atorg.apache.hadoop.yarn.server.resourcemanager.amlauncher.AMLauncher.run(AMLauncher.java:249)/ z+ v$ o( g) j* p. G
        atjava.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
        atjava.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
        atjava.lang.Thread.run(Thread.java:722)
. Failing the application.6 N  G( N1 f9 l& K) z
14/04/29 02:45:07 INFO mapreduce.Job: Counters:0
& M0 e9 z6 h: a7 O
问题原因：namenode,datanode时间同步问题8 y$ c& L  f2 W. h/ q( v& K
) a$ K$ ~1 q3 R0 D5 e
解决办法：多个datanode与namenode进行时间同步，在每台服务器执行：ntpdatetime.nist.gov，确认时间同步成功。6 q' N6 }4 N8 u
最好在每台服务器的/etc/crontab中加入一行：
0 2 * * * root ntpdate time.nist.gov && hwclock –w

3，错误：java.net.SocketTimeoutException: 480000 millistimeout while waiting for channel to be ready for write4 y# J0 W+ W- l#F  o% A4 I
2014-05-06 14:28:09,386 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode:hadoop-datanode1:50010ataXceiver error processing READ_BLOCKoperation  src: /192.168.1.191:48854 dest: /192.168.1.191:50010
java.net.SocketTimeoutException: 480000 millistimeout while waiting for channel to be ready for write. ch :java.nio.channels.SocketChannel[connected local=/192.168.1.191:50010remote=/192.168.1.191:48854]$i  w. V4 w9 o& p
        atorg.apache.hadoop.net.SocketIOWithTimeout.waitForIO(SocketIOWithTimeout.java:246)  Y$ @: l5 Z9 y  L
        atorg.apache.hadoop.net.SocketOutputStream.waitForWritable(SocketOutputStream.java:172)- s$ K8 W) R; |
        atorg.apache.hadoop.net.SocketOutputStream.transferToFully(SocketOutputStream.java:220)
        atorg.apache.hadoop.hdfs.server.datanode.BlockSender.sendPacket(BlockSender.java:546)
        atorg.apache.hadoop.hdfs.server.datanode.BlockSender.sendBlock(BlockSender.java:710)+ N# ]! t7 W8 Q& y
        atorg.apache.hadoop.hdfs.server.datanode.DataXceiver.readBlock(DataXceiver.java:340)
        atorg.apache.hadoop.hdfs.protocol.datatransfer.Receiver.opReadBlock(Receiver.java:101): y, ~, W* P) K* \1 J
        atorg.apache.hadoop.hdfs.protocol.datatransfer.Receiver.processOp(Receiver.java:65)
        atorg.apache.hadoop.hdfs.server.datanode.DataXceiver.run(DataXceiver.java:221)1 q- a' m3 A5 f# `  o8 ^
        atjava.lang.Thread.run(Thread.java:722)

原因：IO超时

解决方法：4 @3 A( S/ l3 z5 g
修改hadoop配置文件hdfs-site.xml，增加dfs.datanode.socket.write.timeout和dfs.socket.timeout两个属性的设置。0 M9 a' R; h9 J; T' U& u
    0 ]2 l+ T8 }3 ]
       dfs.datanode.socket.write.timeout" |0 C% k: u1 q. o9 R. O
       6000000' p  F! j6 g& x! J  U
   

    6 h- q6 G) k/ Z$ S  g" V
       dfs.socket.timeout1 Z3 N1 Q6 A: g
       6000000
   

注意:超时上限值以毫秒为单位。0表示无限制。

4，错误：DataXceiver error processing WRITE_BLOCKoperation
2014-05-0615:21:30,378 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode:hadoop-datanode1:50010ataXceivererror processing WRITE_BLOCK operation  src: /192.168.1.193:34147dest: /192.168.1.191:500100 d3 F/ x) v" t- d/ `1 V' f
java.io.IOException: Premature EOF from inputStream; T4 W) h) A0 Z( o; g
        at org.apache.hadoop.io.IOUtils.readFully(IOUtils.java:194)1 E7 q" J6 V& D2 ]3 J8 u5 r# Y
        atorg.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.doReadFully(PacketReceiver.java:213)9 B/ `& K4 u3 |" N" o; }
        atorg.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.doRead(PacketReceiver.java:134), X4 i5 y8 ?, E, F7 a
        atorg.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.receiveNextPacket(PacketReceiver.java:109)% M0 O4 f& k. H/ H. \/ a2 j0 I
        atorg.apache.hadoop.hdfs.server.datanode.BlockReceiver.receivePacket(BlockReceiver.java:435)
        atorg.apache.hadoop.hdfs.server.datanode.BlockReceiver.receiveBlock(BlockReceiver.java:693)
        atorg.apache.hadoop.hdfs.server.datanode.DataXceiver.writeBlock(DataXceiver.java:569)
        at org.apache.hadoop.hdfs.protocol.datatransfer.Receiver.opWriteBlock(Receiver.java:115)
        atorg.apache.hadoop.hdfs.protocol.datatransfer.Receiver.processOp(Receiver.java:68)
        atorg.apache.hadoop.hdfs.server.datanode.DataXceiver.run(DataXceiver.java:221); Q& ~$ U/ I' q& ?2 _
        at java.lang.Thread.run(Thread.java:722)

原因：文件操作超租期，实际上就是data stream操作过程中文件被删掉了。; v! @( [( o# r# j4 s
$ p/ h6 \+ {9 E+ e
解决办法：
修改hdfs-site.xml（针对2.x版本，1.x版本属性名应该是：dfs.datanode.max.xcievers）：
 
        dfs.datanode.max.transfer.threads % B  |& `$ |. p, S) j7 M; s7o
        8192   q  G4 `1 b$ |

拷贝到各datanode节点并重启datanode即可" X% ?) P& ^. N* z* \( x

5，错误：java.io.IOException: Failed to replace a baddatanode on the existing pipeline due to no more good datanodes being availableto try.& C. g# q% y9 a3 j0 o* T! ~
2014-05-07 12:21:41,820 WARN [Thread-115]org.apache.hadoop.mapreduce.v2.app.MRAppMaster: Graceful stop failed 
org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.io.IOException:Failed to replace a bad datanode on the existing pipeline due to no more gooddatanodes being available to try. (Nodes: current=[192.168.1.191:50010,192.168.1.192:50010], original=[192.168.1.191:50010, 192.168.1.192:50010]). Thecurrent failed datanode replacement policy is DEFAULT, and a client mayconfigure this via 'dfs.client.block.write.replace-datanode-on-failure.policy'in its configuration.
        atorg.apache.hadoop.mapreduce.jobhistory.JobHistoryEventHandler.handleEvent(JobHistoryEventHandler.java:514)
        atorg.apache.hadoop.mapreduce.jobhistory.JobHistoryEventHandler.serviceStop(JobHistoryEventHandler.java:332)
        atorg.apache.hadoop.service.AbstractService.stop(AbstractService.java:221)
        atorg.apache.hadoop.service.ServiceOperations.stop(ServiceOperations.java:52)
        atorg.apache.hadoop.service.ServiceOperations.stopQuietly(ServiceOperations.java:80)
        atorg.apache.hadoop.service.CompositeService.stop(CompositeService.java:159)3 f5 l5 i& y7 P6 k2 _
        atorg.apache.hadoop.service.CompositeService.serviceStop(CompositeService.java:132)
        atorg.apache.hadoop.service.AbstractService.stop(AbstractService.java:221)
        atorg.apache.hadoop.mapreduce.v2.app.MRAppMaster.shutDownJob(MRAppMaster.java:548)
        atorg.apache.hadoop.mapreduce.v2.app.MRAppMaster$JobFinishEventHandler$1.run(MRAppMaster.java:599)
Caused by: java.io.IOException: Failed to replace a bad datanode on theexisting pipeline due to no more good datanodes being available to try. (Nodes:current=[192.168.1.191:50010, 192.168.1.192:50010],original=[192.168.1.191:50010, 192.168.1.192:50010]). The current faileddatanode replacement policy is DEFAULT, and a client may configure this via'dfs.client.block.write.replace-datanode-on-failure.policy' in itsconfiguration.: x$ H! L! r3 f(f1 i/ X
        atorg.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.findNewDatanode(DFSOutputStream.java:860)% ]. B/ u8 e0 c" Q5 f1 H* G# e' {5 m
        at org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.addDatanode2ExistingPipeline(DFSOutputStream.java:925)
        atorg.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.setupPipelineForAppendOrRecovery(DFSOutputStream.java:1031)8 {2 E( W2 G$ C7 O# Y( G2 x
        atorg.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.processDatanodeError(DFSOutputStream.java:823)4 k' U# @0 S2 N3 i2 \6 }) M
        atorg.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:475)1 @/ v; |$ V: [: |
4 ?! g4 J.G  A2 @" V
原因：无法写入；我的环境中有3个datanode，备份数量设置的是3。在写操作时，它会在pipeline中写3个机器。默认replace-datanode-on-failure.policy是DEFAULT,如果系统中的datanode大于等于3，它会找另外一个datanode来拷贝。目前机器只有3台，因此只要一台datanode出问题，就一直无法写入成功。
, `& P. v+ `+r+ B
解决办法：修改hdfs-site.xml文件，添加或者修改如下两项：3 I, |& y9 ]4 G4 C
7 [( K( ?, x# X5 J4 d
  dfs.client.block.write.replace-datanode-on-failure.enable. x9 t. ~7 V: X2 e* v& I1 K! [8 s
  true  M  s: [- Y2 B3 Y/ B
: b0 I" E  g7 T/ C7 }2 U8m& x

  dfs.client.block.write.replace-datanode-on-failure.policy
  NEVER
2 M: v3 v/ @% M# o( N% H8 @
' g; f# V. {% K(D7 [
对于dfs.client.block.write.replace-datanode-on-failure.enable，客户端在写失败的时候，是否使用更换策略，默认是true没有问题。
对于，dfs.client.block.write.replace-datanode-on-failure.policy，default在3个或以上备份的时候，是会尝试更换结点尝试写入datanode。而在两个备份的时候，不更换datanode，直接开始写。对于3个datanode的集群，只要一个节点没响应写入就会出问题，所以可以关掉。

6，错误：org.apache.hadoop.util.DiskChecker$DiskErrorException:Could not find any valid local directory for & a&A1 p; W2 _
14/05/08 18:24:59 INFO mapreduce.Job: Task Id :attempt_1399539856880_0016_m_000029_2, Status : FAILED
Error:org.apache.hadoop.util.DiskChecker$DiskErrorException: Could not find any validlocal directory for attempt_1399539856880_0016_m_000029_2_spill_0.out% w1 N0 b0 d0 a
        atorg.apache.hadoop.fs.LocalDirAllocator$AllocatorPerContext.getLocalPathForWrite(LocalDirAllocator.java:398)
        atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:150)0 k4 o$ [+ |# a+ i7 L" h
        atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:131)
        atorg.apache.hadoop.mapred.YarnOutputFiles.getSpillFileForWrite(YarnOutputFiles.java:159)! J; G1 l( Q9 U9 G1 [
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1573)
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.flush(MapTask.java:1467)
        atorg.apache.hadoop.mapred.MapTask$NewOutputCollector.close(MapTask.java:699)
        at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:769)' f; ~, o! _2 P, ]7 U1 S
        atorg.apache.hadoop.mapred.MapTask.run(MapTask.java:339)
        atorg.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:162)3 |5 g3 ^4 `6 b0 @3 D: E7 Y% T, h
        atjava.security.AccessController.doPrivileged(Native Method)
        atjavax.security.auth.Subject.doAs(Subject.java:415)( \  j3 A# O% V; Q& l
        atorg.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1491)3 h# X2 ^7 }5 ]1 |1 l# l+ V
        atorg.apache.hadoop.mapred.YarnChild.main(YarnChild.java:157)8 S/ k! o( }- S1 o' ~

Container killed by the ApplicationMaster.
/ _, p, U7 _. ]% p' o% o
原因：两种可能，hadoop.tmp.dir或者data目录存储空间不足。# {' [% V  `+ y# I
! X- V4 n* ~+ W# _8 ^$ N$ L! c
解决办法：看了一下我的dfs状态，data使用率不到40%，所以推测是hadoop.tmp.dir空间不足，导致无法创建Jog临时文件。查看core-site.xml发现没有配置hadoop.tmp.dir，因此使用的是默认的/tmp目录，在这目录一旦服务器重启数据就会丢失，因此需要修改。添加：
/ q3 J# X# R2 }, b1 P/ y
hadoop.tmp.dir! ~9 m" j  l- k
/data/tmp
7 L# h' K+ E$ J* y
然后重新格式化：hadoopnamenode -format( K' i+ ?2 b2 E1 N, w% I([7 C
重启。

7.

2014-06-19 10:00:32,181 INFO[org.apache.hadoop.mapred.MapTask] - Ignoring exception during close for org.apache.hadoop.mapred.MapTask$NewOutputCollector@17bda0f2
java.io.IOException: Spill failed
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.checkSpillException(MapTask.java:1540)
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.flush(MapTask.java:1447)
        atorg.apache.hadoop.mapred.MapTask$NewOutputCollector.close(MapTask.java:699)
        atorg.apache.hadoop.mapred.MapTask.closeQuietly(MapTask.java:1997)% \7 v, X5 B0 h5 F. S- Z
        atorg.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:773)6 Z' M  @' E) c6 Z, {0 {: X2 p4 }$ T
        atorg.apache.hadoop.mapred.MapTask.run(MapTask.java:339)
        atorg.apache.hadoop.mapred.LocalJobRunner$Job$MapTaskRunnable.run(LocalJobRunner.java:235)- P" x+ `$ ]; U. D
        atjava.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
        atjava.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:334)
        atjava.util.concurrent.FutureTask.run(FutureTask.java:166)
        atjava.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1110)+ p* S) F2 D7 i  d
        atjava.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:603)& ~3 x' y6 G* M9 t& o: `
        atjava.lang.Thread.run(Thread.java:722)! D: U  ?5 T% a* R
Caused by:org.apache.hadoop.util.DiskChecker$DiskErrorException: Could not find any validlocal directory for output/spill0.out
        atorg.apache.hadoop.fs.LocalDirAllocator$AllocatorPerContext.getLocalPathForWrite(LocalDirAllocator.java:398). M5 S# U+ }5 r0 y9 M6 ]
        atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:150)  u& M4 P3 Z# o* q' ^
        atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:131)8 H% H! A2 L; s8 [8 N, {* J
        atorg.apache.hadoop.mapred.MROutputFiles.getSpillFileForWrite(MROutputFiles.java:146)% Y8 ]/ U  ?9 ^: J. E( w0 {7 f$ ]/ \
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1573)3 U4 \; G2 h" k) |- v2 e1 P
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.access$900(MapTask.java:852)
        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer$SpillThread.run(MapTask.java:1510)" @; u3 b+ P8 ~9 K$ e& d$ \' s
7 C% H3 i  H: Q5 i
( w! T2 d6 z  j# h; j; U
错误原因：本地磁盘空间不足非hdfs (我是在myeclipse中调试程序，本地tmp目录占满)
解决办法：清理、增加空间。

8，

2014-06-23 10:21:01,479 INFO [IPC Server handler 3 on 45207]org.apache.hadoop.mapred.TaskAttemptListenerImpl: Progress of TaskAttemptattempt_1403488126955_0002_m_000000_0 is : 0.308017162014-06-23 10:21:01,512FATAL [IPC Server handler 2 on 45207]org.apache.hadoop.mapred.TaskAttemptListenerImpl: Task:attempt_1403488126955_0002_m_000000_0 - exited : java.io.IOException: Spillfailed        at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.checkSpillException(MapTask.java:1540)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.collect(MapTask.java:1063)       atorg.apache.hadoop.mapred.MapTask$NewOutputCollector.write(MapTask.java:691)       at org.apache.hadoop.mapreduce.task.TaskInputOutputContextImpl.write(TaskInputOutputContextImpl.java:89)       atorg.apache.hadoop.mapreduce.lib.map.WrappedMapper$Context.write(WrappedMapper.java:112)       at com.mediadc.hadoop.MediaIndex$SecondMapper.map(MediaIndex.java:180)       atcom.mediadc.hadoop.MediaIndex$SecondMapper.map(MediaIndex.java:1)       at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:145)       atorg.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:763)       at org.apache.hadoop.mapred.MapTask.run(MapTask.java:339)       atorg.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:162)       at java.security.AccessController.doPrivileged(NativeMethod)        at javax.security.auth.Subject.doAs(Subject.java:415)       at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1491)       atorg.apache.hadoop.mapred.YarnChild.main(YarnChild.java:157)Caused by:org.apache.hadoop.util.DiskChecker$DiskErrorException: Could not find any validlocal directory for attempt_1403488126955_0002_m_000000_0_spill_53.out       atorg.apache.hadoop.fs.LocalDirAllocator$AllocatorPerContext.getLocalPathForWrite(LocalDirAllocator.java:398)       at org.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:150)       atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:131)       atorg.apache.hadoop.mapred.YarnOutputFiles.getSpillFileForWrite(YarnOutputFiles.java:159)       at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1573)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.access$900(MapTask.java:852)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer$SpillThread.run(MapTask.java:1510)2014-06-2310:21:01,513 INFO [IPC Server handler 2 on 45207]org.apache.hadoop.mapred.TaskAttemptListenerImpl: Diagnostics report fromattempt_1403488126955_0002_m_000000_0: Error: java.io.IOException: Spillfailed        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.checkSpillException(MapTask.java:1540)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.collect(MapTask.java:1063)       atorg.apache.hadoop.mapred.MapTask$NewOutputCollector.write(MapTask.java:691)       at org.apache.hadoop.mapreduce.task.TaskInputOutputContextImpl.write(TaskInputOutputContextImpl.java:89)       atorg.apache.hadoop.mapreduce.lib.map.WrappedMapper$Context.write(WrappedMapper.java:112)       atcom.mediadc.hadoop.MediaIndex$SecondMapper.map(MediaIndex.java:180)       at com.mediadc.hadoop.MediaIndex$SecondMapper.map(MediaIndex.java:1)       atorg.apache.hadoop.mapreduce.Mapper.run(Mapper.java:145)       at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:763)       at org.apache.hadoop.mapred.MapTask.run(MapTask.java:339)       at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:162)       at java.security.AccessController.doPrivileged(NativeMethod)        atjavax.security.auth.Subject.doAs(Subject.java:415)       at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1491)       atorg.apache.hadoop.mapred.YarnChild.main(YarnChild.java:157)Caused by:org.apache.hadoop.util.DiskChecker$DiskErrorException: Could not find any validlocal directory for attempt_1403488126955_0002_m_000000_0_spill_53.out       atorg.apache.hadoop.fs.LocalDirAllocator$AllocatorPerContext.getLocalPathForWrite(LocalDirAllocator.java:398)       atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:150)       at org.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:131)       atorg.apache.hadoop.mapred.YarnOutputFiles.getSpillFileForWrite(YarnOutputFiles.java:159)       at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1573)       at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.access$900(MapTask.java:852)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer$SpillThread.run(MapTask.java:1510)2014-06-2310:21:01,514 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl:Diagnostics report from attempt_1403488126955_0002_m_000000_0: Error:java.io.IOException: Spill failed        atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.checkSpillException(MapTask.java:1540)       at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.collect(MapTask.java:1063)       atorg.apache.hadoop.mapred.MapTask$NewOutputCollector.write(MapTask.java:691)       at org.apache.hadoop.mapreduce.task.TaskInputOutputContextImpl.write(TaskInputOutputContextImpl.java:89)       atorg.apache.hadoop.mapreduce.lib.map.WrappedMapper$Context.write(WrappedMapper.java:112)       atcom.mediadc.hadoop.MediaIndex$SecondMapper.map(MediaIndex.java:180)       at com.mediadc.hadoop.MediaIndex$SecondMapper.map(MediaIndex.java:1)       atorg.apache.hadoop.mapreduce.Mapper.run(Mapper.java:145)       at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:763)       at org.apache.hadoop.mapred.MapTask.run(MapTask.java:339)       at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:162)       at java.security.AccessController.doPrivileged(NativeMethod)        atjavax.security.auth.Subject.doAs(Subject.java:415)       at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1491)       at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:157)Causedby: org.apache.hadoop.util.DiskChecker$DiskErrorException: Could not find anyvalid local directory forattempt_1403488126955_0002_m_000000_0_spill_53.out       at org.apache.hadoop.fs.LocalDirAllocator$AllocatorPerContext.getLocalPathForWrite(LocalDirAllocator.java:398)       atorg.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:150)       at org.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:131)       atorg.apache.hadoop.mapred.YarnOutputFiles.getSpillFileForWrite(YarnOutputFiles.java:159)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1573)       at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.access$900(MapTask.java:852)       atorg.apache.hadoop.mapred.MapTask$MapOutputBuffer$SpillThread.run(MapTask.java:1510)2014-06-2310:21:01,516 INFO [AsyncDispatcher event handler]org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl:attempt_1403488126955_0002_m_000000_0 TaskAttempt Transitioned from RUNNING toFAIL_CONTAINER_CLEANUP. C3 s* `3i& q) Q) }
错误很明显，磁盘空间不足，但郁闷的是，进各节点查看，磁盘空间使用不到40%，还有很多空间。4 k2 L+ q9 z6 }4 I6 L% Z+ J' J

郁闷很长时间才发现，原来有个map任务运行时输出比较多，运行出错前，硬盘空间一路飙升，直到100%不够时报错。随后任务执行失败，释放空间，把任务分配给其它节点。正因为空间被释放，因此虽然报空间不足的错误，但查看当时磁盘还有很多剩余空间。+ y1 Q8 m; ]' i4 I- G

这个问题告诉我们，运行过程中的监控很重要。

9.

2015-04-07 23:12:39,837 INFOorg.apache.hadoop.metrics2.impl.MetricsSystemImpl: NameNode metrics systemshutdown complete.) U7 f5 l1 t4 V8 {# j, d
2015-04-07 23:12:39,838 FATAL org.apache.hadoop.hdfs.server.namenode.NameNode:Exception in namenode join
java.io.IOException: There appears to be a gap in the edit log.  Weexpected txid 1, but got txid 41.
        at org.apache.hadoop.hdfs.server.namenode.MetaRecoveryContext.editLogLoaderPrompt(MetaRecoveryContext.java:94)
        atorg.apache.hadoop.hdfs.server.namenode.FSEditLogLoader.loadEditRecords(FSEditLogLoader.java:184)/ Z" l4 r% E  J0 W. l
        at org.apache.hadoop.hdfs.server.namenode.FSEditLogLoader.loadFSEdits(FSEditLogLoader.java:112)
        atorg.apache.hadoop.hdfs.server.namenode.FSImage.loadEdits(FSImage.java:733)
        atorg.apache.hadoop.hdfs.server.namenode.FSImage.loadFSImage(FSImage.java:647)& r) i: B2 }9 N  n/ f( j
        atorg.apache.hadoop.hdfs.server.namenode.FSImage.recoverTransitionRead(FSImage.java:264); e) Q' g7 p6 \# J+ E
        atorg.apache.hadoop.hdfs.server.namenode.FSNamesystem.loadFSImage(FSNamesystem.java:787)& ~; _) G5 k, a3 {  j
        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.loadFromDisk(FSNamesystem.java:568)
        atorg.apache.hadoop.hdfs.server.namenode.NameNode.loadNamesystem(NameNode.java:443)
        atorg.apache.hadoop.hdfs.server.namenode.NameNode.initialize(NameNode.java:491): z/ b: y3 U- l3 u( s* C3 e6 u
        atorg.apache.hadoop.hdfs.server.namenode.NameNode.(NameNode.java:684)6 @+ C5 O* E  f' o. G+J  }
        atorg.apache.hadoop.hdfs.server.namenode.NameNode.(NameNode.java:669)
        at org.apache.hadoop.hdfs.server.namenode.NameNode.createNameNode(NameNode.java:1254)
        atorg.apache.hadoop.hdfs.server.namenode.NameNode.main(NameNode.java:1320)
2015-04-07 23:12:39,842 INFO org.apache.hadoop.util.ExitUtil: Exiting withstatus 11 U1 G8 g/ H- ](b9 q

原因：namenode元数据被破坏，需要修复
解决：恢复一下namenode6 H4 k1 t0 t  P* a4 ?+ j
         hadoop namenode -recover0 O: L! R/ r" Z2 R# V. i) w% q! v
         一路选择c，一般就OK了