This is a simple standalone tool that emulates the Unix 'tree' command in HDFS.
To use it just copy the tool, add it to your path it's ready to use, it has 3 options:
- -p: Specifies the root path. This is mandatory.
- -l: Specifies that you want to use the local filesystem instead of HDFS:
- -s: Specifies that you want to see the size (in human readable format) next to each directory.
- -d: Specifies the maximum depth when displayin the tree.
hdfstree -p / This will print all the HDFS tree. This will take some time... hdfstree -p /user/marc/ Print all the directories starting from /user/marc hdfstree -l -p /Users/marc/Music -s Print all the directories in the /Users/marc/Music directory in the local filesystem and display the sizes.
The output is pretty much like the unix tree command. Have an example:
└──/Users/marc/work/hdfstree ├──.git | ├──branches | ├──hooks | ├──info | ├──logs | | └──refs | | ├──heads | | | └──mdepalol | | └──remotes | | └──origin | ├──objects ...
BUILDING & SOURCE
Build is done with gradle. To create a ready to use script with the libraries just use:
You'll get the executable script + the libraries in
just add this directory (hdfstree/build/install/hdfstree/bin) to your PATH variable and you are done, also, make sure you have the HADOOP_HOME environment variable set. The program will complain if you don't have it.