Tracy R Reed wrote:
> David Looney wrote:
> 
>> Anyone out there know of *any* way to predictably (or at least
>> probably) flush a reiserfs journal of recent transactions ?
> 
> 
> What is the problem you are trying to solve?

Short Answer: Seeking some way to cheat and overwrite data on a
resiserfs partition.

Long,long Answer:  I set up several computers (work, home desktops,
laptops) to use encfs encrypted directories in my home directory, which
in turn resided on a reiserfs "/" partition, thinking I would keep
everything sensitive there.

But, it turned out there were unexpected (by naive me, anyway) snoops,
which transfered information between the "secure" directories and my
home dir.  I found a lot of interesting stuff (for s in `find . -type f
-print`; do grep -l "[0-9][0-9][0-9][ -][0-9][0-9][ -][0-9][0-9][0-9]
[0-9]" $s; done) in the .viminfo directory in my home directory, and
more stuff in the beagle cache directory in my home directory (sneaky
little program, there, that Gnome has introduced).

Oops.

O.K., so I'm resigned to encrypting at least /home (with symlinks for
/tmp and perhaps /var into home), but have been cringing at the thought
of blowing away whole systems (take off reiserfs, wipe partition,
reinstall system with separate partition for /home) to overwrite stuff
already written to disk.

Wouldn't it be nice, I lazily thought, if I could just make a big file
out of all the available disk space, and use shred or some other
file-wipe utility to overwrite it ?

However, the info I found on shred warned against using them on reiserfs
(or other journalling fs) partitions.  I was assuming this was because
of the journal itself, that old information was cached in the journal,
and thinking that if it was possible to "flush" the journal (or turn off
journaling), perhaps this simple strategy might still be viable.

But looking more carefully, it looks like the journal is not the only
problem, in that information may not be written to the same place on
disk even when writing to a specific offset in a file, i.e. a write
does not reliably produce an overwrite.

David Looney

-- 
Good judgment comes from experience, and a lot of that comes from bad
judgment. - Will Rogers


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to