Hadoop complete installation directory structure:
The more important packages are the following 4:
src
hadoop source package. The core code directory is core, hdfs and mapred, they respectively implement the three most important modules of hadoop: basic public library, HDFS implementation and MapReduce implementation
conf
The directory where the configuration file is located. Hadoop has many configuration files, and its design principles can be summarized as the following two points:
1 Make it as modular as possible, that is, each important module has its own configuration file, which makes maintenance and management easy.
2 Separation of dynamic and static, that is, to strip out the configuration options that can be dynamically loaded to form an independent configuration file. For example, before hadoop 1.0.0, configuration options related to job queue permission management were placed in the configuration file mapred-site.xml, which cannot be loaded dynamically, and MapReduce must be restarted after each modification. But starting from version 1.0.0, these configuration options are stripped and put into a separate configuration file mapred-queue-acls.xml, which can be dynamically loaded by hadoop commands.
lib
The third-party packages that the hadoop runtime depends on, including compiled jars and dynamic libraries generated by other languages. These libraries are automatically loaded when hadoop starts or when a user submits a job.
bin
shell startup scripts folder.