26th August
2009
I’ve been wiping a lot of hard drives recently. I use shred to do the job.
shred -n6 -z -v /dev/sdx
What do you do to your drives before disposing of them?
14 Comments
Leave a Reply
ABOUT
a system administrators mutterings
RECENT POSTS
18th January 2012
15th December 2011
05th December 2011
18th November 2011
10th September 2011
Tag Cloud
bash
centos
cleanup
command line
console
cpanel
debian
dns
email
fail
fix
flashback
gentoo
git
gnu screen
google
howto
linux
logging
monitoring
mysql
nagios
nginx
performance analysis
proxy
python
reset
rinse
sar
screen
security
ssh
sysadmin
sysstat
terminal
tip
ubuntu
video
vim
vmware
xen
XenServer
zabbix
zenoss
zenpack
Identica
Twitter
LinkedIn
dd if=/dev/zero of=/dev/whatever BS=1G
relatively quick and nearly guaranteed to be secure.
Yeah that shred command basically does 6 passes of random writes to the device then comes back and writes a single pass of zeros. Maybe I am a bit paranoid. But then again just because you are paranoid doesn’t mean they aren’t out to get you.
I’m sure the program works well, but whenever I think about programs promising to do important things with random numbers, I think of this
http://www.dilbert.com/fast/2001-10-25/
ROFL. Nice
I use DBAN – http://www.dban.org
(Great blog, BTW)
Thanks Joe, I hope you keep coming around!
mmm wonder if the isp used dban on the wrong drive? :/
source: http://www.dban.org (Aug 26th, 11:05am CST, 2009)
My company uses DBAN, and it’s good enough that they use it on old drives from their POS register systems.
dd if=/dev/urandom of=/dev/sdX ; dd if=/dev/zero of=/dev/sdX ;
Even that’s overkill. Either one is more than adequate.
I just repartition, reformat, and reuse. Unless it’s dead, it gets used somewhere. If it’s dead, it gets physically destroyed.
Always a nice read in this context:
http://hostjury.com/blog/view/195/the-great-zero-challenge-remains-unaccepted (I couldn’t find a better version, sorry.)
I haven’t seen anything that indicates that a single ‘dd’ isn’t adequate for any normal disposal process, or that nulls are worse than bytes.
The error that you need to compensate for isn’t that dd might leave recoverable data, it’s that you might forget to dd a drive.
So the ideal process would be to have person ‘a’ dd each drive, with nulls, and have person ‘b’ cat /dev/sdx back to the terminal and make sure that no bytes show up. (i.e. person ‘b’ needs to check that person ‘a’ ran dd.)
A simple tracking system (an ‘x’ mark on the drive) ensures that ‘a’ and ‘b’ both did their jobs.
I have never seen anyone recover from writing over a sector either. Unfortunately I am in no position to dictate corporate policy so I just look for a tool that makes it as easy as possible.
@Jinks
Nice link, I had forgotten about the challenge.
Another vote for DBAN.
It is not a matter of reading from an overwritten sector. It literally takes low level hardware methods to retrieve data from “slightly” overwritten hard drive sectors. The OS reads the last thing written…. With that in mind, the “dd” vs. “shred” argument comes down to what you have on the hard drive and whether someone who could get it actually wants the info….