HADOOP format namenode node preparation script
In general, Live nodes are 0 because the clusterID numbers in namenode and datanode are different, and the incentive is because of repeated formatting. If you don't need to save the data and redo it directly, you need to do the following steps.
ssh hd1 rm /home/hadoop/namenode/* -rf
ssh hd1 rm /home/hadoop/hdfs/* -rf
ssh hd2 rm /home/hadoop/hdfs/* -rf
ssh hd3 rm /home/hadoop/hdfs/* -rf
ssh hd4 rm /home/hadoop/hdfs/* -rf
ssh hd1 rm /home/hadoop/tmp/* -rf
ssh hd2 rm /home/hadoop/tmp/* -rf
ssh hd3 rm /home/hadoop/tmp/* -rf
ssh hd4 rm /home/hadoop/tmp/* -rf
ssh hd1 rm /home/hadoop/hadoop/logs/* -rf
ssh hd2 rm /home/hadoop/hadoop/logs/* -rf
ssh hd3 rm /home/hadoop/hadoop/logs/* -rf
ssh hd4 rm /home/hadoop/hadoop/logs/* -rf
When it's all done, it can be formatted.
hadoop namenode -format
