当前位置: 首页>>代码示例>>Java>>正文


Java BlockInfo.getDatanode方法代码示例

本文整理汇总了Java中org.apache.hadoop.hdfs.server.namenode.BlocksMap.BlockInfo.getDatanode方法的典型用法代码示例。如果您正苦于以下问题:Java BlockInfo.getDatanode方法的具体用法?Java BlockInfo.getDatanode怎么用?Java BlockInfo.getDatanode使用的例子?那么恭喜您, 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.namenode.BlocksMap.BlockInfo的用法示例。


在下文中一共展示了BlockInfo.getDatanode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: testBlockInfoPlaceUpdate

import org.apache.hadoop.hdfs.server.namenode.BlocksMap.BlockInfo; //导入方法依赖的package包/类
@Test
public void testBlockInfoPlaceUpdate() throws IOException {
  insertBlocks(100, true);

  for (Block b : blockList) {
    BlockInfo oldBlock = map.getBlockInfo(b);
    
    // get current locations
    DatanodeDescriptor loc0 = oldBlock.getDatanode(0);
    DatanodeDescriptor loc1 = oldBlock.getDatanode(1);
    assertNotNull(loc0);
    assertNotNull(loc1);
    assertEquals(2, oldBlock.numNodes());

    // prepare new block with different size and GS
    long newGS = b.getGenerationStamp() + 1;
    b.setGenerationStamp(newGS);
    b.setNumBytes(b.getNumBytes() + 10);
    BlockInfo newBlock = new BlockInfo(b, iNode.getReplication());

    LOG.info("Updating block: " + oldBlock + " to: " + newBlock);
    // do update
    newBlock = map.updateINode(oldBlock, newBlock, iNode, 
        iNode.getReplication(), false);
    // preserved new generation stamp
    assertEquals(map.getStoredBlockWithoutMatchingGS(newBlock)
        .getGenerationStamp(), newGS);
    // check locations
    assertEquals(0, newBlock.numNodes());

    // when id is mismatched, the block should not be updated
    newBlock.setBlockId(newBlock.getBlockId() + 1);
    try {
      map.updateINode(oldBlock, newBlock, iNode, iNode.getReplication(), false);
      fail("Should fail here");
    } catch (IOException e) {
      LOG.info("Can't update " + oldBlock + " to: " + newBlock + " "
          + e.getMessage());
    }
  }
}
 
开发者ID:rhli,项目名称:hadoop-EAR,代码行数:42,代码来源:TestBlocksMap.java

示例2: directoryDataNodeUsage

import org.apache.hadoop.hdfs.server.namenode.BlocksMap.BlockInfo; //导入方法依赖的package包/类
/**
 * Iterates through files in the directory dir, and counts the number
 * of usages of datanodes for the files in the directory dir.
 * Just iterates through threshold number of blocks.
 */
private HashMap<DatanodeDescriptor, Integer> directoryDataNodeUsage(INodeDirectory dir,
    int threshold) {
  HashMap<DatanodeDescriptor, Integer> dataNodeUsage = new HashMap<DatanodeDescriptor, Integer>();

  List<INode> children;
  nameSystem.readLock();
  try {
    if (dir.getChildrenRaw() == null) {
      return dataNodeUsage;
    }
    children = new ArrayList<INode>(dir.getChildrenRaw());
    Collections.shuffle(children);

    for (INode node : children) {
      if (!(node instanceof INodeFile)) { // The condition is always false.
        continue;
      }

      INodeFile file = (INodeFile) node;
      BlockInfo[] blocks = file.getBlocks();

      for (BlockInfo block : blocks) {
        if (threshold == 0) {
          return dataNodeUsage;
        }

        int replication = block.numNodes();
        for (int i = 0; i < replication; i++) {
          DatanodeDescriptor datanode = block.getDatanode(i);
          Integer currentUsage = dataNodeUsage.get(datanode);
          dataNodeUsage.put(datanode, currentUsage == null ? 1 : currentUsage + 1);
        }
        threshold--;
      }
    }
  } finally {
    nameSystem.readUnlock();
  }
  return dataNodeUsage;
}
 
开发者ID:rhli,项目名称:hadoop-EAR,代码行数:46,代码来源:BlockPlacementPolicyHBase.java


注:本文中的org.apache.hadoop.hdfs.server.namenode.BlocksMap.BlockInfo.getDatanode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。