On Wed, Jan 27, 2010 at 2:29 AM, Will Stephenson <wstephenson@suse.de> wrote:
On Wednesday 27 January 2010 02:12:58 Bob Smits wrote:
On January 24, 2010 12:10:40 am Dotan Cohen wrote: <snip> One of the things that really irked me about the move from KDE3 to KDE4 was the loss of the secure file erasure facility in Kgpg, with no replacement in sight.
I'm gradually getting used to running KDE4 on one machine while keeping everything else as KDE3 for the time being, and for the most part, KDE4 is quite tolerable. Some parts are really great, but having a gui secure file eraser is quite missing.
User-level Shred was always a "fake sense of security" tool. Explanation here:
http://www.krusader.org/handbook/basic.html
Will
Will, I think the utility of shred is far better than that link states and if there are users that would like to use it, it should be re-instated. My critique of the link you sent:
"Moderns file systems use journalisation."
Journals in general don't have the file data in them, they contain file metadata. (Yes, ext3 now has data journaling support, but it is by no means a true statement to infer all journals include copies of file data.)
"But keep in mind that if you want to be 100% shure that it's impossible that someone can read a deleted file, you need to destroy your harddrive hardware ..."
If that is an effort to say that once data is written to a specific sector, it is forever recoverable via laboratory means (advanced magnetic microscopes), then it is a false statement in all likelihood. We can never know what some extraterrestrial being can recover from a disk drive, but 21st century humans are restricted to the known sciences. Even the USA NIST claims that a sector on a 20GB or denser drive overwritten with a single pass of data makes the previous data unrecoverable via laboratory means. Unfortunately the basis of their statement is a reference to a NSA (National Security Agency) document that is not available to the public. (FOI request was rejected iirc.)
""
I see nothing about SSDs / Flash. Now this is where there is an extremely valid argument that shred is useless. Many of these devices dynamically assign a new erase block every time there is a write operation. Thus the old data is left in a unallocated erase block, even though you think you just overwrote it. Not good. Fortunately, it seems that some SSDs will immediately put that unallocated erase block into an internal erase queue so within a few milliseconds / seconds it does get erased. Unfortunately some don't seem to do that, so there is no easy way for the user to know if a shred operation is safe on a SSD / flash drive. My theory for them agrees with the link you posted. (ie. Use a hammer) Grreg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org