Since the du command used when the capacity is full is difficult to use, I tried wrapping it with ruby

Conclusion

I made a script that wraps the du command with ruby.

Create the name ʻexpand.rb`, paste the script below, and execute it as follows.

# ruby expand.rb (The path you want to find)
$ ruby expand_du.rb /var/lib/
    7GB /var/lib/
    3GB /var/lib/jenkins
    2GB /var/lib/docker
    1GB /var/lib/mysql
    ...
    4KB /var/lib/mysql-files
du_-k_--max-depth_1_var_lib__20200516231438.Saved in log

Scripts (github)

#How to use
# ruby expand_du.rb path
#Example: ruby expand_du.rb ~
#
#effect
# -Execute the du command for the path
# -Output execution result to standard output and file
#
#argument
# -path:Target path to execute the du command

require 'rbconfig'

def main(target_path)
  return puts "ArgumentError:Please put the path as an argument" if target_path.nil?

  max_depth_option_str = if os == :macosx
    "-d"
  else
    "--max-depth"
  end

  exec_command = "du -k #{max_depth_option_str} 1 #{target_path}"
  du_result_str = `#{exec_command}`

  return if du_result_str.empty?

  output_disksizes(du_result_str, exec_command)
end

def output_disksizes(du_result_str, exec_command)
  disk_usages = du_result_str
                  .split("\n")
                  .map{|du_result| DiskUsage.new(du_result)}
                  .sort{|x, y| x.size <=> y.size}.reverse

  output_filename = "#{exec_command.gsub(/( |\/){1,}/, "_")}_#{Time.new.strftime("%Y%m%d%H%M%S")}.log"
  output_file = File.open(output_filename, "w")

  disk_usages.each do |disk_usage|
    puts disk_usage.to_s
    output_file.puts(disk_usage.to_s)
  end

  output_file.close
  puts "#{output_filename}Saved in"
end

class DiskUsage
  attr_reader :size, :path
  def initialize(du_result_line)
    du_result_params = du_result_line.split(" ").map(&:strip)
    @size = du_result_params[0].to_i
    @humanreadable_size, @humanreadable_unit = calc_humanreadable_size
    @path = du_result_params[1]
  end

  def to_s
    #NOTE Specify 5 digits for the time being. Increase when needed
    "#{sprintf("%5d" % @humanreadable_size)}#{@humanreadable_unit} #{@path}"
  end

  def humanreadable_size_with_unit
    "#{@humanreadable_size}#{@humanreadable_unit}"
  end

  private

  def calc_humanreadable_size
    return [@size, "KB"] if mb_size < 1
    return [mb_size, "MB"] if gb_size < 1
    return [gb_size, "GB"] if tb_size < 1
    [tb_size, "TB"]
  end

  def kb_size
    @size
  end

  def mb_size
    kb_size.fdiv(1024).round(1)
  end

  def gb_size
    mb_size.fdiv(1024).round(1)
  end

  def tb_size
    gb_size.fdiv(1024).round(1)
  end
end

def os
  case RbConfig::CONFIG['host_os']
  when /mswin|msys|mingw|cygwin|bccwin|wince|emc/
    :windows
  when /darwin|mac os/
    :macosx
  when /linux/
    :linux
  when /solaris|bsd/
    :unix
  else
    :unknown
  end
end

Background

When your rental server or Mac is full, you'll want to know which folder is the cause.

Therefore, if you are using a Mac, you can see the storage usage status, and even if you are using a rental server, you can see how much you are using with the df command. However, I can only understand the whole thing and not know which folder is the cause.

Therefore, the du command can be used by those who can use the terminal (the ls command is also a candidate, but the ls command knows the size of the file directly under it, but not how big the folder is).

However, this du command is a bit unsatisfactory, probably because it is a UNIX command.

Sorting is not good

To make it easier to read, I use the sort command with the -h option to sort the sizes in descending order, but then 5KB comes above 4GB. I can't help because I'm only looking at the numbers.

It's a hassle to hit twice to record

If you use du, it will output the result to the standard output, but after you do, you will want to record the result. You can output the file from the beginning, but it is troublesome to output the file and then cat. If you devise a command, it will be output to both a file and standard output, but when you need the du command, you are usually in a hurry, so I wanted to do it easily.

Script features

The script that solved the problem above is the first script posted.

The features are as follows.

――It sorts after considering the unit --It outputs both standard output and file (& outputs with a file name that shows which folder was searched)

I made it completely for myself, but if anyone is having trouble with the same thing, please use it.