我正在尝试设置 Hadoop 多节点集群。

当我启动我的集群时，这里是我在控制台中的响应..

hduser@hadoop-master:/usr/local/hadoop$ /usr/local/hadoop/sbin/start-dfs.sh
Starting namenodes on [hadoop-master]
hadoop-master: starting namenode, logging to /usr/local/hadoop/logs/hadoop-hduser-namenode-hadoop-master.out
hadoop-master: starting datanode, logging to /usr/local/hadoop/logs/hadoop-hduser-datanode-hadoop-master.out
hadoop-child: starting datanode, logging to /usr/local/hadoop/logs/hadoop-hduser-datanode-hadoop-child.out
Starting secondary namenodes [0.0.0.0]
0.0.0.0: starting secondarynamenode, logging to /usr/local/hadoop/logs/hadoop-hduser-secondarynamenode-hadoop-master.out
hduser@hadoop-master:/usr/local/hadoop$ jps
21079 NameNode
21258 DataNode
21479 SecondaryNameNode
21600 Jps
hduser@hadoop-master:/usr/local/hadoop$ /usr/local/hadoop/sbin/start-yarn.sh
starting yarn daemons
starting resourcemanager, logging to /usr/local/hadoop/logs/yarn-hduser-resourcemanager-hadoop-master.out
hadoop-child: starting nodemanager, logging to /usr/local/hadoop/logs/yarn-hduser-nodemanager-hadoop-child.out
hadoop-master: starting nodemanager, logging to /usr/local/hadoop/logs/yarn-hduser-nodemanager-hadoop-master.out
hduser@hadoop-master:/usr/local/hadoop$ jps
21079 NameNode
21258 DataNode
22117 Jps
21815 NodeManager
21479 SecondaryNameNode
21658 ResourceManager

你可以看到一个数据节点正在 hadoop-child 机器上启动。

现在，当我尝试获取所有节点信息时。我没有看到显示的所有节点。

hduser@hadoop-master:/usr/local/hadoop$ bin/hdfs dfsadmin -report
Configured Capacity: 21103243264 (19.65 GB)
Present Capacity: 17825124352 (16.60 GB)
DFS Remaining: 17821085696 (16.60 GB)
DFS Used: 4038656 (3.85 MB)
DFS Used%: 0.02%
Under replicated blocks: 0
Blocks with corrupt replicas: 0
Missing blocks: 0
Missing blocks (with replication factor 1): 0

-------------------------------------------------
Live datanodes (1):

Name: 127.0.0.1:50010 (localhost)
Hostname: localhost
Decommission Status : Normal
Configured Capacity: 21103243264 (19.65 GB)
DFS Used: 4038656 (3.85 MB)
Non DFS Used: 3278118912 (3.05 GB)
DFS Remaining: 17821085696 (16.60 GB)
DFS Used%: 0.02%
DFS Remaining%: 84.45%
Configured Cache Capacity: 0 (0 B)
Cache Used: 0 (0 B)
Cache Remaining: 0 (0 B)
Cache Used%: 100.00%
Cache Remaining%: 0.00%
Xceivers: 1
Last contact: Sun Feb 26 17:13:04 UTC 2017

这里需要查看master和child的信息。我有 1 个主节点和 1 个子节点。

********************修复********

根据 Frank 的说法，这是修复它的方式..

1. 编辑/etc/config 文件并给主子 IP 地址。

改变了下面两行

127.0.0.1 localhost hadoop-master
961.118.98.183 hadoop-child

到(在两个节点中)

127.0.0.1 localhost 
961.118.99.251 hadoop-master
961.118.98.183 hadoop-child

2.按以下顺序重启集群..应该重新格式化数据节点。

format will remove only the meta, the datanode's data directories will still be using the old namenode's identity which wil cause the datanode to fail(so delete the directories).

Can you please this order.. 
1) stop the cluster 
2) rm -rf /path/to/datanode/data/dir (in both nodes) 
3) hadoop namenode -format
4) START cluste

最佳答案

设置多节点环境时，必须明确定义namenode 和resourcemanager 地址。

将此属性添加到两个节点中的 core-site.xml，

<property>
  <name>fs.defaultFS</name>
  <value>hdfs://hadoop-master:8020</value>
</property>

在 yarn-site.xml 中也是如此，

<property>
  <name>yarn.resourcemanager.hostname</name>
  <value>hadoop-master</value>
</property>

确保在所有节点的/etc/hosts 文件中，完成了IP 地址和主机名映射。

注意:如果服务正在运行，请停止它们并添加这些属性，然后再重新启动它们。

关于hadoop - Hadoop 集群节点数，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/42471625/