1996-05-04 - Re: encrypted Unix backup software

Header Data

From: VaX#n8 <vax@linkdead.paranoia.com>
To: Bill Stewart <stewarts@ix.netcom.com>
Message Hash: 58b4062b3ac82da1d9b2a36004cc0aa2c1e106906bf7da03654d5a766f9f7f32
Message ID: <199605031253.HAA09103@linkdead.paranoia.com>
Reply To: <199605030425.XAA05924@primus.paranoia.com>
UTC Datetime: 1996-05-04 15:33:58 UTC
Raw Date: Sat, 4 May 1996 23:33:58 +0800

Raw message

From: VaX#n8 <vax@linkdead.paranoia.com>
Date: Sat, 4 May 1996 23:33:58 +0800
To: Bill Stewart <stewarts@ix.netcom.com>
Subject: Re: encrypted Unix backup software
In-Reply-To: <199605030425.XAA05924@primus.paranoia.com>
Message-ID: <199605031253.HAA09103@linkdead.paranoia.com>
MIME-Version: 1.0
Content-Type: text/plain


In message <199605030425.XAA05924@primus.paranoia.com>, Bill Stewart writes:
>>Although I could probably hack up "catblock" to do the job, and use
>>a line of the form
>>dump -0uBf ... | symmetric_cipher | catblock blockfactor > /dev/tape

>Yup [agrees on cmdline]
>You don't need to write "catblock", though - there's the "dd" command
>designed for just such applications...
>Newer versions may handle multiple tapes a bit better.

Hmm, I've got this a few times from people and I just wanted to clear
up a few points.
1) BSD dd doesn't treat EOT specially:
   /* ... If a partial write, and it's a character device, just warn.
          If a tape device, quit. ... */
   GNU dd (from what I remember) doesn't, either.
2) "dd" will only work safely across all tape types and all size pipes
   using the degenerate form "dd ibs=1 obs=(your_bf_here) conv=sync ...".

I'll hack up catblock today.  Who says we don't code :)





Thread