On 14/11/2025 12:54, Ricky Tigg wrote:
Hello. Finding out the algorithm the shred tool relies on for its shredding
operation will need more than reading the 'info coreutils shred'
documentation. Unusual to be put in position to investigate here in order
to get that information. Neither 'shred' 's '-v' invocation leads to
communicate that.
$ echo 01 > test
$ shred -v -f -n 1 -u -z test
shred: test: pass 1/2 (random)...
shred: test: pass 2/2 (000000)...
shred: test: removing
shred: test: renamed to 0000
shred: 0000: renamed to 000
shred: 000: renamed to 00
shred: 00: renamed to 0
shred: test: removed
Yet this is essential information for the user in order to consider
entrusting any resource whatsoever to a program for processing.
Yes, --verbose should give more info on what it outputs.
I've updated its description in man page and info docs at:
https://github.com/coreutils/coreutils/commit/faf8d8998
Out-of scope in regard to the subject
The target file to overwrite is written on Btrfs file system which design
does not satisfy the program assumption - file system and hardware
overwrite data in place. Btrfs being a copy-on-write that incorporates
features of both log-structured and journal-operating, I therefore suppose
is must be currently configured with those features on Fedora. Neither
should I invoke '-u' and '--remove' since the file is written on /home,
itself mounted on block device /dev/sd<_>.
Yes, shred really is only appropriate when used directly with devices
these days, and even then there may be issues due to ware levelling etc.
This is mentioned in the man page, and detailed in the info page already.
thanks,
Padraig
