Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 9 May 1998 15:41:36 -0700 (PDT)
From:      Tom <tom@sdf.com>
To:        Terry Lambert <tlambert@primenet.com>
Cc:        freebsd-hackers@FreeBSD.ORG
Subject:   Re: dump/restore problem (was: Network problem with 2.2.6-RELEASE)
Message-ID:  <Pine.BSF.3.95q.980509152620.10035A-100000@misery.sdf.com>
In-Reply-To: <199805092206.PAA28038@usr07.primenet.com>

next in thread | previous in thread | raw e-mail | index | archive | help

On Sat, 9 May 1998, Terry Lambert wrote:

> > > The real question here is whether or not you can eliminate the IDE
> > > as a cause of the problem.
> > 
> >   The real question is why have you determined that the cause is IDE and
> > forced me to disprove it?  There is no evidence of file corruption or
> > other damage that would occur if IDE is broken.
> 
> Because it works for me, but it doesn't work for you.
> 
> Because the dump/restore software has not mysteriously mutated.
> 
> Because IDE vs. SCSI is the most obvious difference in our configurations
> besides the tape drive, which you have already eliminated by using a
> disk file.

  There are several other obvious differences:

- 9GB filesystem (yours) vs 32GB filesystem (mine)
- average file size is about 1.5GB (since I expect that you use your
system for devel, not archiving, you'd be way under that)

> >   The next real question, is if your really serious about debugging this
> > problem, why you don't want to see the first few megs of one of these
> > damaged dump archives?
> 
> I'd be happy to look at it if you put it up for FTP some place.

  I can't do that.

> My mailbox would not accept something that large, so putting it up
> for FTP is the only way you are going to be able to communicate it to me

  Since the first few megs are all headers, they would compress well,
probably resulting in a less than 5MB attachment.  I'm pretty sure
primenet can handle that (even AOL accepts e-mail up to 3MB). 

> > > I *know* that John has done a lot of work on the IDE driver; much
> > > of this work is in -current, not -stable, and I doubt it has been
> > > seriously tested with all possible hardware.
> > 
> >   I wish you could develop the ability to read PRs, and look up bin/4683
> > Many different versions of FreeBSD are affected.
> 
> This is irrelevent.  The relevent portion is my doubt that dump/restore
> has been tested with all possible hardware.
> 
> What other software are you aware of that utilizes the raw disk
> device, such that you think your data path has been tested?

  What about newfs, and fsck?

> >   Also, the IDE drive is very solid.  Thousands of FreeBSD users use
> > IDE.  John has been working on new functionality for wdc (UltraDMA
> > mainly).
> 
> Let's test this theory.
> 
> Let's see, what would show a difference between your IDE and my SCSI?

  I have other small IDE systems that work ok (1GB filesystem to 2GB DAT).

  The biggest difference between your system and mine, is that I'm using a
32GB filesystem.

...
> >   I don't have 20 to 30GB of free SCSI disk space.  Remember, this is a
> > 32GB filesystem generally about 70 to 90% full.
> 
> But you claim that it repeats at a much smaller size.  I believe you
> were claiming 4G.

  Nope, never claimed that.  I've never had less than 15GB of data on this
filesystem at any one time.

Tom


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95q.980509152620.10035A-100000>