After spending an afternoon installing Hadoop from scratch, I was excited to start it!
Error: JAVA_HOME is not set and could not be found.
He told me that JAVA_HOME was not found?
I obviously installed java, why can't I find it?
I checked java -version and found that there is no problem.
Solution:
Later I found that the java path on Hadoop was not configured, that is, the Hadoop environment variables were not configured.
Enter the hadhoop folder.
cd hadoop/etc/hadoop/
vim hadoop-env.sh
G跳到文末
addition
export JAVA_HOME=/usr/java/jdk1.8.0_181-amd64 每个人版本号不同而且这个地址要到bin目录的上一层才可以！
export HADOOP_CONF_DIR=/usr/local/hadoop/etc/hadoop
Esc :wq 保存退出
source hadoop-env.sh 重新执行sh文件 就成功了
Finally, fill in a pit: It is recommended not to modify the jdk name.
After I modified it, the name changed. After looking at the detailed information, wls said that the soft link had changed, and then the entire folder became unusable. 
In fact, there is no impact if you directly delete or ignore these two folders (no impact here means that you can start Hadoop) , but they are really ugly and I have a little obsessive-compulsive disorder . I directly forcefully deleted the java folder and wanted to reinstall it. java environment, the result is the following article .