The program is a companion to the "du" file system utility. It is intended for the scenario where a system admininstrator discovers with "df" that there is no more or little extra space left on the file system. In a company with many departments the sysadmin cannot be expected to know what is causing the file system to fill up. The sysadmin may run the program to output a list of the n largest items in a given set of paths and thereafter take action on them.
The program first quicksorts its arguments to remove duplicates. Next it processes each argument in turn, descending into subdirectories. Directory sizes are computed recursively from the sizes of the entries in that directory. During this process the largest items are collected in a list. The program outputs the list at the end. There is a human-readable output format where the sizes are formatted as kilobytes, megabytes etc.
dumx(1) dumx(1) NAME dumx - displays the largest files and/or directories in a given list of paths SYNOPSIS dumx [OPTION]... [FILE]... DESCRIPTION This program will recursively process all files and direc- tories given as arguments and output the largest items, where directory sizes are computed by summing content sizes. OPTIONS -n ENTRIES Set the number of entries to retrieve, default is 10. -h Produce human-readable output using suffixes K, M, G etc. on file sizes. -v Invert, i.e. find smallest items. -type ARG Files/directories only, the corresponding values of ARG are 'f' and 'd'. -xdev Don't descend into other file systems. -uid NUMBER Restrict to a given user id. -gid NUMBER Restrict to a given group id. -user NAME Restrict to a given user name. -group NAME Restrict to a given group name. -regex PATTERN Restrict to files matching extended POSIX regex pattern. -iregex PATTERN Like -regex, but case-insensitive. AUTHOR Marko Riedel, firstname.lastname@example.org SEE ALSO du(1) dumx(1)