使用hdfs上传文件报错org.apache.hadoop.hdfs.CannotObtainBlockLengthExceptio

Question

问题遇到的现象和发生背景

使用hdfs上传文件报错，如果是空文件上传就会成功，如果是有内容的文件就会上传失败

运行结果及报错内容

org.apache.hadoop.hdfs.CannotObtainBlockLengthException: Cannot obtain block length for LocatedBlock{BP-293563276-192.168.0.11-1634095772355:blk_1073784416_43592; getBlockSize()=0; corrupt=false; offset=0; locs=[DatanodeInfoWithStorage[192.168.0.11:9866,DS-0f59c3c4-7aae-4baa-b7df-786930725c03,DISK]]}
at org.apache.hadoop.hdfs.DFSInputStream.readBlockLength(DFSInputStream.java:363) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DFSInputStream.fetchLocatedBlocksAndGetLastBlockLength(DFSInputStream.java:270) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DFSInputStream.openInfo(DFSInputStream.java:201) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DFSInputStream.(DFSInputStream.java:185) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DFSClient.openInternal(DFSClient.java:1048) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DFSClient.open(DFSClient.java:1011) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DistributedFileSystem$4.doCall(DistributedFileSystem.java:319) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DistributedFileSystem$4.doCall(DistributedFileSystem.java:315) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.fs.FileSystemLinkResolver.resolve(FileSystemLinkResolver.java:81) ~[hadoop-common-3.2.1.jar!/:?]
at org.apache.hadoop.hdfs.DistributedFileSystem.open(DistributedFileSystem.java:327) ~[hadoop-hdfs-client-3.2.1.jar!/:?]
at org.apache.hadoop.fs.FileSystem.open(FileSystem.java:899) ~[hadoop-common-3.2.1.jar!/:?]
at com.jf.common.oss.service.impl.HdfsOSSImpl.getObject(HdfsOSSImpl.java:68) [jfcommon-oss-2.1.0.jar!/:?]
at com.jf.common.oss.JfOSSClient.getObject(JfOSSClient.java:90) [jfcommon-oss-2.1.0.jar!/:?]
at com.jfh.oss.OSSClientProxy.getOSSObject(OSSClientProxy.java:98) [classes!/:?]
at com.jfh.util.MergePartHelper.mergePartFile(MergePartHelper.java:78) [classes!/:?]
at com.jfh.helper.ViewFileHelper.mergePartFile(ViewFileHelper.java:125) [classes!/:?]
at com.jfh.controller.JFZGFileController.mergePartFile(JFZGFileController.java:99) [classes!/:?]
at com.jfh.controller.JFZGFileController$$FastClassBySpringCGLIB$$2bf96aa4.invoke() [classes!/:?]
at org.springframework.cglib.proxy.MethodProxy.invoke(MethodProxy.java:218) [spring-core-5.1.15.RELEASE.jar!/:5.1.15.RELEASE]
at org.springframework.aop.framework.CglibAopProxy$CglibMethodInvocation.invokeJoinpoint(CglibAopProxy.java:752) [spring-aop-5.1.15.RELEASE.jar!/:5.1.15.RELEASE]
at org.springframework.aop.framework.ReflectiveMethodInvocation.proceed(ReflectiveMethodInvocation.java:163) [spring-aop-5.1.15.RELEASE.jar!/:5.1.15.RELEASE]

我的解答思路和尝试过的方法

使用 hdfs fsck -openforwrite这个命令得到的数据中没有openforwrite

Answer 1

信息不足不是很好判断，找了下源码，报错的方法如下：

/** Read the block length from one of the datanodes. */
  private long readBlockLength(LocatedBlock locatedblock) throws IOException {
    assert locatedblock != null : "LocatedBlock cannot be null";
    int replicaNotFoundCount = locatedblock.getLocations().length;

    final DfsClientConf conf = dfsClient.getConf();
    final int timeout = conf.getSocketTimeout();
    LinkedList<DatanodeInfo> nodeList = new LinkedList<DatanodeInfo>(
        Arrays.asList(locatedblock.getLocations()));
    LinkedList<DatanodeInfo> retryList = new LinkedList<DatanodeInfo>();
    boolean isRetry = false;
    StopWatch sw = new StopWatch();
    while (nodeList.size() > 0) {
      DatanodeInfo datanode = nodeList.pop();
      ClientDatanodeProtocol cdp = null;
      try {
        cdp = DFSUtilClient.createClientDatanodeProtocolProxy(datanode,
            dfsClient.getConfiguration(), timeout,
            conf.isConnectToDnViaHostname(), locatedblock);

        final long n = cdp.getReplicaVisibleLength(locatedblock.getBlock());

        if (n >= 0) {
          return n;
        }
      } catch (IOException ioe) {
        checkInterrupted(ioe);
        if (ioe instanceof RemoteException) {
          if (((RemoteException) ioe).unwrapRemoteException() instanceof
              ReplicaNotFoundException) {
            // replica is not on the DN. We will treat it as 0 length
            // if no one actually has a replica.
            replicaNotFoundCount--;
          } else if (((RemoteException) ioe).unwrapRemoteException() instanceof
              RetriableException) {
            // add to the list to be retried if necessary.
            retryList.add(datanode);
          }
        }
        DFSClient.LOG.debug("Failed to getReplicaVisibleLength from datanode {}"
              + " for block {}", datanode, locatedblock.getBlock(), ioe);
      } finally {
        if (cdp != null) {
          RPC.stopProxy(cdp);
        }
      }

      // Ran out of nodes, but there are retriable nodes.
      if (nodeList.size() == 0 && retryList.size() > 0) {
        nodeList.addAll(retryList);
        retryList.clear();
        isRetry = true;
      }

      if (isRetry) {
        // start the stop watch if not already running.
        if (!sw.isRunning()) {
          sw.start();
        }
        try {
          Thread.sleep(500); // delay between retries.
        } catch (InterruptedException e) {
          Thread.currentThread().interrupt();
          throw new InterruptedIOException(
              "Interrupted while getting the length.");
        }
      }

      // see if we ran out of retry time
      if (sw.isRunning() && sw.now(TimeUnit.MILLISECONDS) > timeout) {
        break;
      }
    }

    // Namenode told us about these locations, but none know about the replica
    // means that we hit the race between pipeline creation start and end.
    // we require all 3 because some other exception could have happened
    // on a DN that has it.  we want to report that error
    if (replicaNotFoundCount == 0) {
      return 0;
    }

    throw new CannotObtainBlockLengthException(locatedblock);
  }

从报错的代码中，只有两个地方能够返回，此处没有返回而抛了这个异常。而没有返回的原因是获取备份的可见长度n<0。代码如下：

final long n = cdp.getReplicaVisibleLength(locatedblock.getBlock());

        if (n >= 0) {
          return n;
        }

数据备份一般与存储有关，注意到你截图上面集群只有一个数据节点。不知道是不是该数据节点的存储已经用光了，可以检查了看看。
信息不足，只能猜测到这些了。

Answer 2

参考看看：

Cannot obtain block length for LocatedBlock故障分析与解决_mnasd的博客-CSDN博客 在执行hive任务的时候经常会遇到如下错误：Caused by: java.io.IOException: Cannot obtain block length for LocatedBlock{BP-739651031-192.168.1.50-1452872943245:blk_1329315607_256707953; getBlockSize()=6604762; corrupt=fa... https://blog.csdn.net/mnasd/article/details/84560221

Answer 3

重新安装一次吧

Answer 4

看看这边呢
https://www.cnblogs.com/cssdongl/p/6700512.html

Answer 5

如果没有特别重要得数据，重新装也方便，我之前有写过，从Hadoop到spark 到flink都有得，可以借鉴下

Answer 6

这篇文章讲的很详细，请看：Hadoop上传文件到hdfs报错:org.apache.hadoop.ipc.RemoteException(java.io.IOException)