我在单节点集群上安装了 hadoop。当我使用 start=all.sh 启动所有 hadoop 服务时，一切正常。但是，每当我停止所有服务并重新启动 hadoop 服务时，都会出现以下异常，我必须再次重新格式化文件系统。现在我正在开发中，每当我重新格式化时我都会复制文件。但是我不能在生产中出现这种行为。我检查了日志，这里是异常(exception)。我的 etc/hosts 文件也是一样的，我禁用了 IPV6
2012-11-03 18:49:45,542 INFO org.apache.hadoop.hdfs.server.namenode.NameNode: STARTUP_MSG:
/************************************************************
STARTUP_MSG: Starting NameNode
STARTUP_MSG: host = nikhil-VirtualBox/127.0.1.1
STARTUP_MSG: args = []
STARTUP_MSG: version = 1.0.4
STARTUP_MSG: build = https://svn.apache.org/repos/asf/hadoop/common/branches /branch-1.0 -r 1393290; compiled by 'hortonfo' on Wed Oct 3 05:13:58 UTC 2012
************************************************************/
2012-11-03 18:49:45,738 INFO org.apache.hadoop.metrics2.impl.MetricsConfig: loaded properties from hadoop-metrics2.properties
2012-11-03 18:49:45,750 INFO org.apache.hadoop.metrics2.impl.MetricsSourceAdapter: MBean for source MetricsSystem,sub=Stats registered.
2012-11-03 18:49:45,751 INFO org.apache.hadoop.metrics2.impl.MetricsSystemImpl: Scheduled snapshot period at 10 second(s).
2012-11-03 18:49:45,751 INFO org.apache.hadoop.metrics2.impl.MetricsSystemImpl: NameNode metrics system started
2012-11-03 18:49:45,899 INFO org.apache.hadoop.metrics2.impl.MetricsSourceAdapter: MBean for source ugi registered.
2012-11-03 18:49:45,902 WARN org.apache.hadoop.metrics2.impl.MetricsSystemImpl: Source name ugi already exists!
2012-11-03 18:49:45,910 INFO org.apache.hadoop.metrics2.impl.MetricsSourceAdapter: MBean for source jvm registered.
2012-11-03 18:49:45,910 INFO org.apache.hadoop.metrics2.impl.MetricsSourceAdapter: MBean for source NameNode registered.
2012-11-03 18:49:45,932 INFO org.apache.hadoop.hdfs.util.GSet: VM type = 32-bit
2012-11-03 18:49:45,934 INFO org.apache.hadoop.hdfs.util.GSet: 2% max memory = 19.33375 MB
2012-11-03 18:49:45,934 INFO org.apache.hadoop.hdfs.util.GSet: capacity = 2^22 = 4194304 entries
2012-11-03 18:49:45,934 INFO org.apache.hadoop.hdfs.util.GSet: recommended=4194304, actual=4194304
2012-11-03 18:49:46,020 INFO org.apache.hadoop.hdfs.server.namenode.FSNamesystem: fsOwner=hduser
2012-11-03 18:49:46,020 INFO org.apache.hadoop.hdfs.server.namenode.FSNamesystem: supergroup=supergroup
2012-11-03 18:49:46,021 INFO org.apache.hadoop.hdfs.server.namenode.FSNamesystem: isPermissionEnabled=true
2012-11-03 18:49:46,024 INFO org.apache.hadoop.hdfs.server.namenode.FSNamesystem: dfs.block.invalidate.limit=100
2012-11-03 18:49:46,024 INFO org.apache.hadoop.hdfs.server.namenode.FSNamesystem: isAccessTokenEnabled=false accessKeyUpdateInterval=0 min(s), accessTokenLifetime=0 min(s)
2012-11-03 18:49:46,169 INFO org.apache.hadoop.hdfs.server.namenode.FSNamesystem: Registered FSNamesystemStateMBean and NameNodeMXBean
2012-11-03 18:49:46,191 INFO org.apache.hadoop.hdfs.server.namenode.NameNode: Caching file names occuring more than 10 times
2012-11-03 18:49:46,194 INFO org.apache.hadoop.hdfs.server.common.Storage: Storage directory /tmp/hadoop-hduser/dfs/name does not exist.
2012-11-03 18:49:46,196 ERROR org.apache.hadoop.hdfs.server.namenode.FSNamesystem: FSNamesystem initialization failed.
org.apache.hadoop.hdfs.server.common.InconsistentFSStateException: Directory /tmp/hadoop-hduser/dfs/name is in an inconsistent state: storage directory does not exist or is not accessible.
at org.apache.hadoop.hdfs.server.namenode.FSImage.recoverTransitionRead(FSImage.java:303)
at org.apache.hadoop.hdfs.server.namenode.FSDirectory.loadFSImage(FSDirectory.java:100)
at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.initialize(FSNamesystem.java:388)
at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.<init>(FSNamesystem.java:362)
at org.apache.hadoop.hdfs.server.namenode.NameNode.initialize(NameNode.java:276)
at org.apache.hadoop.hdfs.server.namenode.NameNode.<init>(NameNode.java:496)
at org.apache.hadoop.hdfs.server.namenode.NameNode.createNameNode(NameNode.java:1279)
at org.apache.hadoop.hdfs.server.namenode.NameNode.main(NameNode.java:1288)
2012-11-03 18:49:46,197 ERROR org.apache.hadoop.hdfs.server.namenode.NameNode: org.apache.hadoop.hdfs.server.common.InconsistentFSStateException: Directory /tmp/hadoop-hduser/dfs/name is in an inconsistent state: storage directory does not exist or is not accessible.
at org.apache.hadoop.hdfs.server.namenode.FSImage.recoverTransitionRead(FSImage.java:303)
at org.apache.hadoop.hdfs.server.namenode.FSDirectory.loadFSImage(FSDirectory.java:100)
at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.initialize(FSNamesystem.java:388)
at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.<init>(FSNamesystem.java:362)
at org.apache.hadoop.hdfs.server.namenode.NameNode.initialize(NameNode.java:276)
at org.apache.hadoop.hdfs.server.namenode.NameNode.<init>(NameNode.java:496)
at org.apache.hadoop.hdfs.server.namenode.NameNode.createNameNode(NameNode.java:1279)
at org.apache.hadoop.hdfs.server.namenode.NameNode.main(NameNode.java:1288)
最佳答案
不要将 HDFS 安装在 /tmp 目录中。它会在您每次启动时被您的操作系统清除。
我个人已将我的 HDFS 安装在 /srv/hdfs/ 中，但我想这是个人喜好问题。
关于hadoop - 每当我停止和启动 hadoop 服务时都必须反复格式化 HDFS，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/13214499/