Subject: Re: A few mt(1) questions
To: Aaron J. Grier <hauke@Espresso.Rhein-Neckar.DE>
From: Brian A. Seklecki <lavalamp@spiritual-machines.org>
List: netbsd-users
Date: 03/22/2002 04:24:32
On Wed, 20 Mar 2002, Aaron J. Grier wrote:

> On Tue, Mar 19, 2002 at 02:18:25AM -0500, Brian A. Seklecki wrote:
> > On Mon, 18 Mar 2002, Aaron J. Grier wrote:
> >


> as for writing the same bit over and over again, maybe the drive is
> performing compression?  I think you can turn it off with

Okay, I disabled compression on both drives and tried again w/ /dev/random
(or /dev/urandom).  Results were the same over several hours:

/dev/rst0
107776+0 records in
107776+0 records out
55181312 bytes transferred in 16377 secs (3369 bytes/sec)

/dev/rst1
107908+0 records in
107908+0 records out
55248896 bytes transferred in 16388 secs (3371 bytes/sec)

...those numbers are indeed different than w/ /dev/zero.  At first maybe I
_hoped_ that they might come out exactly the same, meaning that dd reaches
a certain threshold and stops incrimenting (but fails to error out).  That
was not the case.

I see another thread on a send-pr/kern related to mt(1) developing.  Is
this at all related? (CCing someone who mentioned the problem affects
Tandburg SLR [X] drives, one of which I'm using: SLR 5)

> "mt -f /dev/nrst0 compress 0."