Subject: Re: A few mt(1) questions
To: Aaron J. Grier <>
From: Brian A. Seklecki <>
List: netbsd-users
Date: 03/19/2002 20:23:38
> dd if=/dev/zero of=/dev/rmt0
> does the trick for me under 1.4 and 1.5 on my pmaxen with TZK10 QIC
> drive.  I could've sworn I also did this on i386 with an exabyte as
> well.  maybe using the raw device is the trick?

That didn't work either.  I let both tapes run on a dd for about the same
time (couple hours).  When I returned, the drives "sounded" like they were
"stuck", writing the last part of the tape over and over again.

$ echo $TAPE
$ dd if=/dev/zero of=$TAPE
^C1076610+0 records in
1076609+0 records out
551223808 bytes transferred in 13418 secs (41080 bytes/sec)

$ echo $TAPE
$ dd if=/dev/zero of=$TAPE
^C1090261+0 records in
1090260+0 records out
558213120 bytes transferred in 13483 secs (41401 bytes/sec)

You would _think_ they'd get to the end of the tape, error, dd would exit
with accurate stats, and the tapes would rewind.

Like for example, when I tar-pipe -> "ssh dd" a backup of my home
directory with the following:

/usr/local/bin/tar cvfBb - 1024 /export/home/lavalamp | /usr/local/bin/ssh
-oProtocol=2 hurt dd of=/dev/rst0 obs=1024b

It completes with.

423936+0 records in
414+0 records out
217055232 bytes transferred in 1739 secs (124816 bytes/sec)

423936 records @ 512 byte blocks is right with:

$ du -ks ~
213347  /export/home/lavalamp


> --