[parisc-linux] i dont know jack

[parisc-linux] i dont know jack

[Rick Vernam]

[-- Attachment #1: Type: text/plain, Size: 497 bytes --]

i work at place that gets tons of computer junk from companies that are getting rid of their equip.  one of our requirements is to kill the partitions on the HDs ( it is known that will no completely remove data, just make is a little more difficult to get too ).  so we are getting HP machines with pa-risc stuff and most wont boot ( i know enough to know how to initiate a boot ).  is there some way, using this parisc-linux that i can quickly kill partitions?

-Rick
rickvernam@hotmail.com

[-- Attachment #2: Type: text/html, Size: 981 bytes --]

Re: [parisc-linux] i dont know jack

[James P. Kinney III]

[-- Attachment #1: Type: text/plain, Size: 1616 bytes --]

During the install process, one must create partitions to format then
install onto. Getting to that point is not too slow. Doing it from any
ol' HP box that comes in the door may be the biggest challenge. 

If you can get one box to run with SCSI connections available for other
drives, boot the working boox with all the drives you want blanked,
fdisk them to one big linux partition, then shutdown. It will require
that you pull the drives, but in batch mode of 6 at a time, it will
likely be faster than 6 boots and partial setups.

Any of that old junk you want to throw my way (Atlanta) would be fun :)
My office already looks like a museum.

On Thu, 2002-07-25 at 11:43, Rick Vernam wrote:
> i work at place that gets tons of computer junk from companies that are getting rid of their equip.  one of our requirements is to kill the partitions on the HDs ( it is known that will no completely remove data, just make is a little more difficult to get too ).  so we are getting HP machines with pa-risc stuff and most wont boot ( i know enough to know how to initiate a boot ).  is there some way, using this parisc-linux that i can quickly kill partitions?
> 
> -Rick
> rickvernam@hotmail.com
-- 
James P. Kinney III   \Changing the mobile computing world/
President and CEO      \          one Linux user         /
Local Net Solutions,LLC \           at a time.          /
770-493-8244             \.___________________________./

GPG ID: 829C6CA7 James P. Kinney III (M.S. Physics)
<jkinney@localnetsolutions.com>
Fingerprint = 3C9E 6366 54FC A3FE BA4D 0659 6190 ADC3 829C 6CA7 



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 232 bytes --]

Re: [parisc-linux] i dont know jack

[Juergen Braukmann]

> Rick Vernam schrieb:
> 
> i work at place that gets tons of computer junk >from companies that
> are getting rid of their equip.  one of our requirements is to kill
> the partitions on the HDs ( it is known that will no completely remove
> data, just make is a little more difficult to get too ).  so we are
> getting HP machines with pa-risc stuff and most wont boot ( i know
> enough to know how to initiate a boot ).  is there some way, using
> this parisc-linux that i can quickly kill partitions?
> 
> -Rick
> rickvernam@hotmail.com

Hi Rick, there probably is. Boot into some nini root environment and
type:

dd if=/dev/null of=/dev/sda [assuming THE drive is sda]

this will overwrite the intire disk with zero bytes (or was /dev/zero
the proper device for that??). Atlernetivly, /dev/random might be a good
source of rubbish data. ;-)

there is the bs= parameter as well (block size). You probably need to
experiment a bit with that, I tried to copy a 20GB disk via dd and it
was dead slow, but I used 8KB as bs. I'd now start with a value of 4-8
MB.

dd if=/dev/null of=/dev/sda bs=512 count=1 will overwrite the boot
sector and partition table; but that's DOS-logic, I do not know how much
of that applies to original HP-UX disks.

Juergen

-- 
===========================================     __   _
Juergen Braukmann juergen.braukmann@gmx.de| -o)/ /  (_)__  __ ____  __
Tel: 0201-743648  dk4jb@db0qs.#nrw.deu.eu | /\\ /__/ / _ \/ // /\ \/ /
===========================================_\_v __/_/_//_/\_,_/ /_/\_\

Re: [parisc-linux] i dont know jack

[Ralf Hildebrandt]

Am 25.07.2002 um 18:29:15 +0200 schrieb Juergen Braukmann folgendes:

> Hi Rick, there probably is. Boot into some nini root environment and
> type:
> 
> dd if=/dev/null of=/dev/sda [assuming THE drive is sda]

bs=1M

Increase the blocksize heftily.

> this will overwrite the intire disk with zero bytes (or was /dev/zero
> the proper device for that??). Atlernetivly, /dev/random might be a good
> source of rubbish data. ;-)

/dev/zero is better.
Don't try /dev/random, it's too slow

-- 
Ralf Hildebrandt (Im Auftrag des Referat V A)   Ralf.Hildebrandt@charite.de
Charite Campus Virchow-Klinikum                 Tel.  +49 (0)30-450 570-155
Referat V A - Kommunikationsnetze -             Fax.  +49 (0)30-450 570-916
Why you can't find your system administrators:
S/He is playing DOOM, to release pent-up tension created by lusers. 

Re: [parisc-linux] i dont know jack

[Sandy Harris]

Juergen Braukmann wrote:

> Hi Rick, there probably is. Boot into some nini root environment and
> type:
> 
> dd if=/dev/null of=/dev/sda [assuming THE drive is sda]
> 
> this will overwrite the intire disk with zero bytes (or was /dev/zero
> the proper device for that??).

/dev/zero gives a stream of null bytes.

Don't use /dev/null; it returns EOF on all reads, or did when I learned
the rules on older Unix and I suspect that still true.

> Atlernetivly, /dev/random might be a good source of rubbish data. ;-)

Not /dev/random. It was designed to produce high-grade random numbers
for critical applications like generating PGP keys. It blocks if you
try to take out more random data than it has input entropy.

/dev/urandom does not block, so you could use that.

It would likely be better to use a little program that just seeds itself
from /dev/urandom and then cranks out lots of psuedo-random crud
quickly.
The FreeS/WAN libraries include source you could use:
http://www.freeswan.org/freeswan_snaps/CURRENT-SNAP/doc/manpage.d/ipsec_prng.3.html
 
> there is the bs= parameter as well (block size). You probably need to
> experiment a bit with that, I tried to copy a 20GB disk via dd and it
> was dead slow, but I used 8KB as bs. I'd now start with a value of 4-8
> MB.

Yes, use a large block size.

How thorough do you need to be? At one extreme, just trashing the
partition table or superbloack may be all you need. At the other,
you may need to do a fair bit of programming.

The classic paper on secure deletion is:
http://www.cs.auckland.ac.nz/~pgut001/pubs/secure_del.html

I once read a US gov't standard for overwriting disks with
non-classified
data on them. (For classified data, you destroy the disk.) It wanted a
minimum of three overwrites, all-0s, all-1s and random data.

The hard part was that you had to guarantee to do that everywhere,
including
blocks the OS or drive had marked bad, things outside partitions that
the OS
couldn't see, ...

A handy loop for cheap but fairly thorough deletions is:

for( i = u = 0 ; i < 4 ; i++, u += 0x55555555 )

This walks each nybble of u through the values 0000, 0101, 1010, 1111 so
you gat the US gov'ts all-0s and all-1s plus a couple of others.

Re: [parisc-linux] i dont know jack

[Alan Cox]

On Thu, 2002-07-25 at 16:43, Rick Vernam wrote:
> i work at place that gets tons of computer junk from companies that are getting rid of their equip.  one of our requirements is to kill the partitions on the HDs ( it is known that will no completely remove data, just make is a little more difficult to get too ).  so we are getting HP machines with pa-risc stuff and most wont boot ( i know enough to know how to initiate a boot ).  is there some way, using this parisc-linux that i can quickly kill partitions?
> 

Well one silly way to do it would be to simply install Linux on the
machines telling it to use the whole of the disks. At that point you can
put "tested" on the disks 8) and "Free Debian Linux" on the marketing 8)

Re: [parisc-linux] i dont know jack

[Bryan W. Headley]

Alan Cox wrote:
> On Thu, 2002-07-25 at 16:43, Rick Vernam wrote:
> 
>>i work at place that gets tons of computer junk from companies that are getting rid of their equip.  one of our requirements is to kill the partitions on the HDs ( it is known that will no completely remove data, just make is a little more difficult to get too ).  so we are getting HP machines with pa-risc stuff and most wont boot ( i know enough to know how to initiate a boot ).  is there some way, using this parisc-linux that i can quickly kill partitions?
>>
> 
> 
> Well one silly way to do it would be to simply install Linux on the
> machines telling it to use the whole of the disks. At that point you can
> put "tested" on the disks 8) and "Free Debian Linux" on the marketing 8)

What, silly? Our admins used to prep Solaris workstations that way. Real 
inexpensive way to differentiate your products ("our Pengiuns pre-hatched")


-- 
____               .:.                 ____
Bryan W. Headley - bwheadley@earthlink.net

RE: [parisc-linux] i dont know jack

[CARSON,KEVIN (HP-Canada,ex1)]

There are a couple of methods (assuming /dev/sda is the SCSI disk you wish
to erase):

1) dd if=/dev/zero of=/dev/sda bs=1M count=1

2) script parted or sfdisk to discover partitions and erase them

3) wipe -f -q /dev/sda

1 is the fastest.  Does little to none data protection.  2 is probably just
too fiddly and error prone.  3 is slow but does a reasonable four pass data
scrub of the whole disk as well as erasing the partitions.

Slow is relative.  It took 10.5 minutes to do an older 2GB drive on a 20
MB/s SCSI channel.  However, one could do these in parallel.  A rule of
thumb (based upon the assumption that recycling depots get old disks on old
SCSI channels) is about four disks in parallel per SCSI channel.  Of course,
this is really slow compared to the dd -- under a second on the same setup.

I can see three strategies for the setup:

A) Boot from a scsi disk to an external connector.

B) Boot from a network connector.

C) Remove the disks from the unit and access them through a testbed.

A is pretty straightforward but it assumes an external connector with
compatible SCSI (SE or differential, etc).  Adding a SCSI adapter is not
always an option as PA-RISC Linux doesn't drive HP-PB cards.

B is more effort to get the environment setup but may result in a simpler
process.  It still suffers from the lack of accessibility for anything
connected to HP-PB for instance.

C is more physical work but should probably always work.  A testbed unit
choice might be an older workstation with both differential and SE SCSI
connectors.  There would be the physical hassle of getting different
connector types hooked up but that is solved with various (custom) cables.

Hope this helps!

Kevin