Conversation

Notices

1. Embed this notice
   feld (feld@bikeshed.party)'s status on Tuesday, 20-Feb-2024 05:59:19 JST feld
   Remember when everyone was like "you have to secure erase your drives / overwrite with random data multiple times with some NIST standard or your data can still be recovered"?
   
   Did anyone ever witness this actually happen: successful data recovery from a drive that was overwritten only once ?
   
   I feel like this is only a plausible scenario for some very very old HDDs
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Tuesday, 20-Feb-2024 10:18:13 JST feld

     in reply to

     • tedu
     @tedu How does it even work though? That part never made sense. Do the bits have a shadow? 😂
   • Embed this notice
     tedu (tedu@honk.tedunangst.com)'s status on Tuesday, 20-Feb-2024 10:18:14 JST tedu

     in reply to

     @feld it's probably been 30 years since anyone got data off a single wiped drive.

     SSD could be a little different, because it's not easy to wipe spare area.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Tuesday, 20-Feb-2024 10:31:21 JST feld

     in reply to

     • tedu
     • knapjack
     @knapjack @tedu oh I meant the spinning rust drives, sorry!
     
     SSDs are very complicated so I just use full disk encryption on everything to be safe. Who knows what's lurking, plus the extra flash they overprovision ...
     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     knapjack (knapjack@bonk.cozysumo.space)'s status on Tuesday, 20-Feb-2024 10:31:23 JST knapjack

     in reply to

     • tedu

     @tedu @feld SSDs have some pros and cons. If folks are using any of the built-in ATA sanitize commands to wipe the drive, or doing a cryptographic erase (which is pretty common these days), extremely safe. Doing a single-pass wipe, in theory you could trick it into reading previously deallocated space and retrieve some blocks, but the odds of either them being successful or the results being useful is super low.

   • Embed this notice
     rocc@misskey.bubbletea.dev's status on Tuesday, 20-Feb-2024 10:48:14 JST Rocc

     in reply to

     @feld@bikeshed.party yes with hdds

     Doughnut Lollipop 【記録係】:blobfoxgooglymlem: likes this.
   • Embed this notice
     tedu (tedu@honk.tedunangst.com)'s status on Tuesday, 20-Feb-2024 12:46:50 JST tedu

     in reply to

     @feld actually... the magnet only polarizes most of the bits, but the ones that aren't are now very small. in the past, the unwritten regions could be large enough to read.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     knapjack (knapjack@bonk.cozysumo.space)'s status on Tuesday, 20-Feb-2024 12:49:06 JST knapjack

     in reply to

     • tedu

     @feld @tedu Oh yeah, spinning disc, I would totally single pass wipe.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Tuesday, 20-Feb-2024 12:52:23 JST Haelwenn /элвэн/ :triskell:

     in reply to
     @feld I think I've only ever done single wipes from /dev/urandom data, at least in cases where I wouldn't forget (these days I just encrypt if it could be sensitive anyway).