Storage monitor

erdtree

There are many disk usage tools, so it requires something special to pull me out of my du dependency stupor. I looked at one called Parallel Disk Usage recently, because it was very quick, and erdtree is just as quick and takes a different UI approach. Where Parallel Disk Usage outputs a bar graph showing usage for every directory, erdtree's erd command will instead collate totals in the same way du does. It also offers many of the same command-line options as du, including a human-readable option and long output, so it's a great drop-in replacement – and it does all this while looking much better than du. The default layout takes inspiration from another perennial command, tree, with ASCII lines branching off from directories to subdirectories. The number of levels is configurable, as is the wonderful use of color and accompanying icons for every node in the output. The tree view can be inverted, or disabled entirely, and looks fantastic in 256 colors.

But it's the configurability of erdtree that makes such a difference, either from the command-line arguments or from the optional configuration file. It will even parse .gitignore files when deciding which files and folders to show, a feature not seen outside of Git itself, and if that isn't enough, regular expressions and glob patterns can be used for file and directory searching and matching. There's just as much control over sorting too, from reverse names to alteration dates, and support for storage values with either binary or SI prefixes. All this is done at incredible speed, thanks to configurable multithreading, and you can scale your CPU usage according to the demands of your filesystem. This might be useful if you're streaming or in a video meeting at the same time, for instance, and solves the biggest problem with du and tree, which was their lack of speed. Either way, creating an alias from du to erd makes a great command-line upgrade.

Project Website

https://github.com/solidiquis/erdtree

The erdtree executable is erd, which can be typed quickly, but not as quickly as it can generate deep-level output after analyzing your storage.

RSS hosting