1. 參考//hadoop.apache.org/docs/r2.10.2/hadoop-project-dist/hadoop-hdfs/HdfsRollingUpgrade.html
2. 軟件準備
1. #新版本準備
   wget //archive.apache.org/dist/hadoop/common/hadoop-2.10.2/hadoop-2.10.2.tar.gz
   tar -zxvf hadoop-2.10.2.tar.gz -C ../program/
   cp ${HADOOP_HOME}/etc/hadoop/*-site.xml ${DIR}/hadoop-2.10.2/etc/hadoop/
   cp ${HADOOP_HOME}/etc/hadoop/slaves ${DIR}/hadoop-2.10.2/etc/hadoop/
   #拷貝到其他機器
   scp -r ${ip}:${DIR}/hadoop-2.10.2 ${DIR}/hadoop-2.10.2
3. 準備滾動升級
1. 運行“ hdfs dfsadmin -rollingUpgrade prepare ”來創建用于回滾的 fsimage。

運行“ hdfs dfsadmin -rollingUpgrade query ”來檢查回滾映像的狀態。等待并重新運行該命令，直到顯示“ Proceed with rolling upgrade ”消息。

1. 升級主備神經網絡
1. 關閉并升級NN2。（如果DN和NN在一臺機子上面DN也要關閉，因為要改環境變量）
   ${HADOOP_HOME}/sbin/hadoop-daemon.sh stop namenode
2. 使用“ -rollingUpgrade started ”選項將NN2作為備用啟動。
   $HADOOP_HOME/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR --script hdfs start namenode -rollingUpgrade started
3. 從NN1故障轉移到NN2 ，以便NN2變為活動狀態，NN1變為備用狀態。
   hdfs haadmin -failover nn2 nn1
4. 關閉并升級NN1。
   ${HADOOP_HOME}/sbin/hadoop-daemon.sh stop namenode 
   使用“ -rollingUpgrade started ”選項將NN1作為備用啟動。
   $HADOOP_HOME/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR --script hdfs start namenode -rollingUpgrade started
2. 注意點
1. 啟動前要切換HADOOP_HOME
2. vim ~/.bash_profile
   export HADOOP_HOME=${DIR}/hadoop-2.10.1
   改成
   export HADOOP_HOME=${DIR}/hadoop-2.10.2
   source ~/.bash_profile
3. 如果遇到no namenode to stop
1. 修改HADOOP_PID_DIR配置
1. mkdir -p ~/hadoop-data/pids
   vim ${HADOOP_HOME}/etc/hadoop/hadoop-env.sh，更改HADOOP_PID_DIR的值
   export HADOOP_PID_DIR=~/hadoop-data/pids
   vim ${HADOOP_HOME}/etc/hadoop/hdfs-site.xml
   dfs.ha.automatic-failover.enabled改成dfs.ha.automatic-failover.enabled.${nameservice的id}
2. 用jps查看現有的hadoop進程，并且kill掉

jps | grep   -E ' NameNode|NodeManager|DataNode|JobHistoryServer|Jps|JournalNode' | awk '{print $1}' | xargs kill

jps | grep   -E ' NodeManager|JournalNode' | awk '{print $1}' | xargs kill

1. 重新啟動原來的hadoop程序

#Start the HDFS NameNode with the following command on the designated node as hdfs:
$HADOOP_HOME/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR --script hdfs start namenode -rollingUpgrade started

#Start a HDFS DataNode with the following command on each designated node as hdfs:
$HADOOP_HOME/sbin/hadoop-daemons.sh --config $HADOOP_CONF_DIR --script hdfs start datanode

#If etc/hadoop/slaves and ssh trusted access is configured (see Single Node Setup), all of the HDFS processes can be started with a utility script. As hdfs:
# $HADOOP_HOME/sbin/start-dfs.sh

#Start the YARN with the following command, run on the designated ResourceManager as yarn:
$HADOOP_HOME/sbin/yarn-daemon.sh --config $HADOOP_CONF_DIR start resourcemanager

#Run a script to start a NodeManager on each designated host as yarn:
$HADOOP_HOME/sbin/yarn-daemons.sh --config $HADOOP_CONF_DIR start nodemanager

#Start a standalone WebAppProxy server. Run on the WebAppProxy server as yarn. If multiple servers are used with load balancing it should be run on each of them:
$HADOOP_HOME/sbin/yarn-daemon.sh --config $HADOOP_CONF_DIR start proxyserver

#If etc/hadoop/slaves and ssh trusted access is configured (see Single Node Setup), all of the YARN processes can be started with a utility script. As yarn:
# $HADOOP_HOME/sbin/start-yarn.sh

#Start the MapReduce JobHistory Server with the following command, run on the designated server as mapred:
$HADOOP_HOME/sbin/mr-jobhistory-daemon.sh --config $HADOOP_CONF_DIR start historyserver

#查看集群狀態
hdfs haadmin -getAllServiceState

1. 如果兩個NN都是standby，就要強制設置主節點
1. hdfs haadmin -transitionToActive --forcemanual nn1
   #${HADOOP_HOME}/bin/hdfs zkfc -formatZK
1. 升級DN
1. 選擇一小部分數據節點（例如特定機架下的所有數據節點）。
1. 運行“ hdfs dfsadmin -shutdownDatanode <DATANODE_HOST:IPC_PORT> upgrade ”以關閉所選數據節點之一。
   運行“ hdfs dfsadmin -getDatanodeInfo <DATANODE_HOST:IPC_PORT> ”檢查并等待數據節點關閉。
   升級并重啟數據節點。
   $HADOOP_HOME/sbin/hadoop-daemons.sh --config $HADOOP_CONF_DIR --script hdfs start datanode
   對子集中所有選定的數據節點并行執行上述步驟。
2. 重復上述步驟，直到集群中的所有datanode都升級完畢。
2. 完成滾動升級
1. 運行“ hdfs dfsadmin -rollingUpgrade finalize ”以完成滾動升級。
3. 啟動hive的元數據$HIVE_HOME/bin/hive --service metastore &