设置多节点hadoop集群块池ID不匹配

Question

在设置多节点 hadoop 集群时，我遇到了几个问题。通过不同的门户网站进行正确设置。出现了一些基本问题
我正在使用 Hadoop 2.8.5 在主从配置中设置一个 2 节点集群。
在第一个机器上使用 hdfs namenode format

格式化 namenode

clusterID 和 BlockpoolID 的分配如下：

#Fri Mar 29 11:14:41 IST 2019
namespaceID=576041649
clusterID=CID-98480e8d-f7a9-4e1a-8997-400a7aa150c3
cTime=1553838281164
storageType=NAME_NODE
blockpoolID=BP-954411427-x.x.x.y-1553838281164
layoutVersion=-63

现在在第 2 台机器上，我运行命令 hdfs namenode format -clusterId CID-98480e8d-f7a9-4e1a-8997-400a7aa150c3

#Fri Mar 29 11:15:38 IST 2019
namespaceID=304822257
clusterID=CID-98480e8d-f7a9-4e1a-8997-400a7aa150c3
cTime=1553838338130
storageType=NAME_NODE
blockpoolID=BP-1421744029-x.x.x.x-1553838338130
layoutVersion=-63

考虑到slave和master应该有相同的clusterID，如有错误请指正。配置似乎工作正常，但我在 logs/hadoop-cassandra-datanode-localnosql1.log 和 logs/hadoop-cassandra-datanode-localnosql2.log

的日志中收到错误

2019-03-29 11:25:44,009 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed for Block pool BP-954411427-x.x.x.y-1553838281164 (Datanode Uuid 4b90bebb-3c34-43d5-8285-6ec6dfefc0a7) service to localnosql1/x.x.x.x:8020 Blockpool ID mismatch: previously connected to Blockpool ID BP-954411427-x.x.x.y-1553838281164 but now connected to Blockpool ID BP-1421744029-x.x.x.x-1553838338130  
2019-03-29 11:25:49,010 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed for Block pool BP-954411427-x.x.x.y-1553838281164 (Datanode Uuid 4b90bebb-3c34-43d5-8285-6ec6dfefc0a7) service to localnosql1/x.x.x.x:8020 Blockpool ID mismatch: previously connected to Blockpool ID BP-954411427-x.x.x.y-1553838281164 but now connected to Blockpool ID BP-1421744029-x.x.x.x-1553838338130  
2019-03-29 11:25:54,012 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed for Block pool BP-954411427-x.x.x.y-1553838281164 (Datanode Uuid 4b90bebb-3c34-43d5-8285-6ec6dfefc0a7) service to localnosql1/x.x.x.x:8020 Blockpool ID mismatch: previously connected to Blockpool ID BP-954411427-x.x.x.y-1553838281164 but now connected to Blockpool ID BP-1421744029-x.x.x.x-1553838338130  
2019-03-29 11:25:59,013 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed for Block pool BP-954411427-x.x.x.y-1553838281164 (Datanode Uuid 4b90bebb-3c34-43d5-8285-6ec6dfefc0a7) service to localnosql1/x.x.x.x:8020 Blockpool ID mismatch: previously connected to Blockpool ID BP-954411427-x.x.x.y-1553838281164 but now connected to Blockpool ID BP-1421744029-x.x.x.x-1553838338130  
2019-03-29 11:26:04,014 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed for Block pool BP-954411427-x.x.x.y-1553838281164 (Datanode Uuid 4b90bebb-3c34-43d5-8285-6ec6dfefc0a7) service to localnosql1/x.x.x.x:8020 Blockpool ID mismatch: previously connected to Blockpool ID BP-954411427-x.x.x.y-1553838281164 but now connected to Blockpool ID BP-1421744029-x.x.x.x-1553838338130

这些错误日志暗示了什么？

所有主从节点上的blockpool ID是否需要像clusterId一样，如果是怎么办？

Answer 1

为什么要格式化 namenode 两次？理想情况下，在多节点配置中，有一个名称节点和许多数据节点。第一次设置时，你用“hdfs namenode -format”初始化namenode，然后启动datanodes，它工作正常。

如果您正在尝试多主配置（同时有多个名称节点运行），我不确定这是否可行。

如果您正在尝试为 namenode 配置主备，您可以尝试以下步骤

Hadoop Namenode HA setup

设置多节点hadoop集群块池ID不匹配

Setting up multinode hadoop cluster Blockpool ID mismatch

hadoop

hadoop2