On Tue 19 Apr 2022 at 07:19:58 (+0200), DdB wrote: > So i came up with the idea to create a sort of inventory using a sparse > copy of empty files only (using mkdir, truncate + touch). The space > requirements are affordable (like 2.3M for an inventory representing > 3.5T of data). The effect being, that find will see those files by > name/directory/time/size and permissions, allowing to find duplicates > according to those attributes quite nicely. > > Since i started doing that, i always do know exactly, what's out there > and where to find it, as if i had only the inodes at hand. Suits MY > needs. :-)
That sounds somewhat like my scheme for indexing my caddies, USB sticks and SD cards. When I unmount them, a script makes three indexes: the first running updatedb to a private database, the second running find to produce a list of strictly alphabetically ordered files (full name, --full-time and size) and the directories containing them, and the third, most like yours, running ls -lAR to another private database. The three indexes are propagated to my other hosts. Apart from being a plain text file, the output of ls -lAR, stored either as ….lslR or ….lslR.gz, can be navigated with Midnight Commander (mc) just as if it were a real filesystem. (Permissions don't concern me.) Cheers, David.
