Path: g2news1.google.com!news1.google.com!news.glorb.com!news.ripco.com!not-for-mail
From: Bruce Esquibel <b...@ripco.com>
Newsgroups: comp.sys.sun.admin
Subject: Re: st.conf for Quantum DAT 160
Date: Thu, 24 Mar 2011 11:30:00 +0000 (UTC)
Organization: Ripco Communications Inc.
Lines: 29
Message-ID: <imf9vo$6ae$1@remote5bge0.ripco.com>
References: <imabvn$bpt$1@news.eternal-september.org> <imcocp$pi9$1@remote5bge0.ripco.com> <imd5t0$c5a$1@news.eternal-september.org>
NNTP-Posting-Host: shell3.ripco.com
X-Trace: remote5bge0.ripco.com 1300966200 6478 74.8.102.118 (24 Mar 2011 11:30:00 GMT)
X-Complaints-To: usenet@ripco.com
NNTP-Posting-Date: Thu, 24 Mar 2011 11:30:00 +0000 (UTC)
User-Agent: tin/1.4.5-20010409 ("One More Nightmare") (UNIX) (SunOS/5.10 (i86pc))

Tristram Scott <tristram.sc...@ntlworld.com> wrote:

> Thanks for your thoughts, Bruce.  Unfortunately, I see it being slower than
> its predecessor both with and without compression, which makes me think
> that compression is not the issue.

> I wondered if it might be a block size issue, but my tests with dd suggest
> that it doesn't care much about blocksize, as long as it is more than a few
> kB.  Maybe it is just slow.


Only other thought I had is maybe something with the termination.

That is a regular scsi device, correct?

Maybe it has built in terminators and you have another terminator on the 
cable. Usually that kills the bus from working at all but maybe it's seeing
enough data to work but erroring out constantly.

How slow is slow anyway?

I don't remember those things being speed demons, might have the numbers
wrong but 15-20GB an hour I think was about it. Took like 3 1/2 to 4 hours
to backup a 100GB file system with 80% utilization.

Maybe that was for the AIT tapes, really don't remember, heh.

-bruce
b...@ripco.com