当前位置: 首页>>代码示例>>Java>>正文


Java DataNode.getDNRegistrationForBP方法代码示例

本文整理汇总了Java中org.apache.hadoop.hdfs.server.datanode.DataNode.getDNRegistrationForBP方法的典型用法代码示例。如果您正苦于以下问题:Java DataNode.getDNRegistrationForBP方法的具体用法?Java DataNode.getDNRegistrationForBP怎么用?Java DataNode.getDNRegistrationForBP使用的例子?那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在org.apache.hadoop.hdfs.server.datanode.DataNode的用法示例。


在下文中一共展示了DataNode.getDNRegistrationForBP方法的2个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于系统推荐出更棒的Java代码示例。

示例1: runTest

import org.apache.hadoop.hdfs.server.datanode.DataNode; //导入方法依赖的package包/类
private static void runTest(final String testCaseName,
                            final boolean createFiles,
                            final int numInitialStorages,
                            final int expectedStoragesAfterTest) throws IOException {
  Configuration conf = new HdfsConfiguration();
  MiniDFSCluster cluster = null;

  try {
    cluster = new MiniDFSCluster
        .Builder(conf)
        .numDataNodes(1)
        .storagesPerDatanode(numInitialStorages)
        .build();
    cluster.waitActive();

    final DataNode dn0 = cluster.getDataNodes().get(0);

    // Ensure NN knows about the storage.
    final DatanodeID dnId = dn0.getDatanodeId();
    final DatanodeDescriptor dnDescriptor =
        cluster.getNamesystem().getBlockManager().getDatanodeManager().getDatanode(dnId);
    assertThat(dnDescriptor.getStorageInfos().length, is(numInitialStorages));

    final String bpid = cluster.getNamesystem().getBlockPoolId();
    final DatanodeRegistration dnReg = dn0.getDNRegistrationForBP(bpid);
    DataNodeTestUtils.triggerBlockReport(dn0);

    if (createFiles) {
      final Path path = new Path("/", testCaseName);
      DFSTestUtil.createFile(
          cluster.getFileSystem(), path, 1024, (short) 1, 0x1BAD5EED);
      DataNodeTestUtils.triggerBlockReport(dn0);
    }

    // Generate a fake StorageReport that is missing one storage.
    final StorageReport reports[] =
        dn0.getFSDataset().getStorageReports(bpid);
    final StorageReport prunedReports[] = new StorageReport[numInitialStorages - 1];
    System.arraycopy(reports, 0, prunedReports, 0, prunedReports.length);

    // Stop the DataNode and send fake heartbeat with missing storage.
    cluster.stopDataNode(0);
    cluster.getNameNodeRpc().sendHeartbeat(dnReg, prunedReports, 0L, 0L, 0, 0,
        0, null);

    // Check that the missing storage was pruned.
    assertThat(dnDescriptor.getStorageInfos().length, is(expectedStoragesAfterTest));
  } finally {
    if (cluster != null) {
      cluster.shutdown();
    }
  }
}
 
开发者ID:naver,项目名称:hadoop,代码行数:54,代码来源:TestNameNodePrunesMissingStorages.java

示例2: runTest

import org.apache.hadoop.hdfs.server.datanode.DataNode; //导入方法依赖的package包/类
private static void runTest(final String testCaseName,
                            final boolean createFiles,
                            final int numInitialStorages,
                            final int expectedStoragesAfterTest) throws IOException {
  Configuration conf = new HdfsConfiguration();
  MiniDFSCluster cluster = null;

  try {
    cluster = new MiniDFSCluster
        .Builder(conf)
        .numDataNodes(1)
        .storagesPerDatanode(numInitialStorages)
        .build();
    cluster.waitActive();

    final DataNode dn0 = cluster.getDataNodes().get(0);

    // Ensure NN knows about the storage.
    final DatanodeID dnId = dn0.getDatanodeId();
    final DatanodeDescriptor dnDescriptor =
        cluster.getNamesystem().getBlockManager().getDatanodeManager().getDatanode(dnId);
    assertThat(dnDescriptor.getStorageInfos().length, is(numInitialStorages));

    final String bpid = cluster.getNamesystem().getBlockPoolId();
    final DatanodeRegistration dnReg = dn0.getDNRegistrationForBP(bpid);
    DataNodeTestUtils.triggerBlockReport(dn0);

    if (createFiles) {
      final Path path = new Path("/", testCaseName);
      DFSTestUtil.createFile(
          cluster.getFileSystem(), path, 1024, (short) 1, 0x1BAD5EED);
      DataNodeTestUtils.triggerBlockReport(dn0);
    }

    // Generate a fake StorageReport that is missing one storage.
    final StorageReport reports[] =
        dn0.getFSDataset().getStorageReports(bpid);
    final StorageReport prunedReports[] = new StorageReport[numInitialStorages - 1];
    System.arraycopy(reports, 0, prunedReports, 0, prunedReports.length);

    // Stop the DataNode and send fake heartbeat with missing storage.
    cluster.stopDataNode(0);
    cluster.getNameNodeRpc().sendHeartbeat(dnReg, prunedReports, 0L, 0L, 0, 0,
        0, null, true);

    // Check that the missing storage was pruned.
    assertThat(dnDescriptor.getStorageInfos().length, is(expectedStoragesAfterTest));
  } finally {
    if (cluster != null) {
      cluster.shutdown();
    }
  }
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:54,代码来源:TestNameNodePrunesMissingStorages.java


注:本文中的org.apache.hadoop.hdfs.server.datanode.DataNode.getDNRegistrationForBP方法示例由纯净天空整理自Github/MSDocs等开源代码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。