Access control for Hadoop

HDFS supports permission control but is weak in support. The design of HDFs is based on POSIX model, which supports read and write execution control by user, user group and other users. Under the Linux command line, you can use the following command to modify the permissions of the file, the file owner, and the group to which the file belongs:
Hadoop fs–chmod (Modify file owner, file belongs to group, other user's read, write, execute permissions)
Haddop Fs–chown (modify file owner)
Hadoop fs–chgrp (modify file group)
Different users can access specific files by using different Linux accounts. Users who launch Hadoop HDFS systems are super users and can do anything.
Note that HDFS itself does not provide a user name, a user group creation, and when the client invokes the Hadoop file Operations Command, Hadoop recognizes the user name and user group of the process executing the command, and then uses this user name and group to check file permissions.
The user name =linux the ' whoami ' in the command, and the group name equals ' bash-c groups '.
According to the above principle, the file system permissions management of HDFs is weak. If the user knows the HDFs name node address and port number, after installing the HDFs client, use the same username as name node to gain access to all files.