Some pits encountered in configuring Hadoop [Ubuntu 16.04, Hadoop 3.1]

Pit one:

The openjdk that comes with Ubuntu has only the most basic operating environment. If you need a development environment, you also need to install jdk

sudo apt-get install openjdk-8-jdk

Then configure the relevant environment variables....

Pit two:

Set up local ssh access without password

    generate key

Enter the command ssh-keygen -t rsa and press Enter all the way

    copy into public key

sudo cp ~/.ssh/id_rsa.pub ~/.ssh/authorized_keys

Pit three:

When running start_dfs.sh, the definitions of several variables are still missing:

For example HDFS_DATANODE_USER

See the online solution, add to the beginning of start_dfs.sh and stop_dfs.sh:

HDFS_DATANODE_USER=hadoop  
HADOOP_SECURE_DN_USER=hdfs  
HDFS_NAMENODE_USER=hadoop  
HDFS_SECONDARYNAMENODE_USER=hadoop

Note, switch hadoop to the username you are currently using