From owner-freebsd-current  Thu Apr 10 03:24:04 1997
Return-Path: <owner-current>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id DAA27503
          for current-outgoing; Thu, 10 Apr 1997 03:24:04 -0700 (PDT)
Received: from utserv.mcc.ac.uk (utserv.mcc.ac.uk [130.88.200.6])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id DAA27433
          for <freebsd-current@freebsd.org>; Thu, 10 Apr 1997 03:23:31 -0700 (PDT)
Received: from albatross.mcc.ac.uk by utserv.mcc.ac.uk with SMTP (PP);
          Thu, 10 Apr 1997 11:22:45 +0100
Received: (from ip@localhost) by albatross.mcc.ac.uk (8.8.5/8.6.12) id LAA15505 
          for freebsd-current@freebsd.org;
          Thu, 10 Apr 1997 11:22:36 +0100 (BST)
From: Ian Pallfreeman <ip@mcc.ac.uk>
Message-Id: <199704101022.LAA15505@albatross.mcc.ac.uk>
Subject: Re: wdx interrupt timeouts (fwd)
To: freebsd-current@freebsd.org
Date: Thu, 10 Apr 1997 11:22:35 +0100 (BST)
Reply-To: ip@mcc.ac.uk
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

I've switched this to the current list, since it's worse than I thought.

Doug White wrote:
> On Tue, 8 Apr 1997, Ian Pallfreeman wrote:
> > Apr  8 15:07:54 lurch /kernel: wd3: interrupt timeout:
> > Apr  8 15:07:54 lurch /kernel: wd3: status 58<rdy,seekdone,drq> error 0
> > [...]
> Do you have IDE spindown enabled?  That's what it looks like happened--
> wd3 spun down and the system tried to poke it, and it took a little too
> long to come back to ready.  Nothing to worry about.

Nothing to do with IDE spindown (had a couple of other folks suggesting that).
This happens under heavy load -- e.g. mirroring the ftp archive.

The problem has become worse over the last few days. First it changed to:

	Apr  9 14:06:48 lurch /kernel: wd3e: reverting to non-multi sector mode reading fsbn 4213504 of 4213456-4213518 (wd3 bn 4213504; cn 4180 tn 1 sn 1)wd3: status 59<rdy,seekdone,drq,err> error 40<uncorr>
	Apr  9 14:06:55 lurch /kernel: wd3e: hard error reading fsbn 4213506 of 4213456-4213518 (wd3 bn 4213506; cn 4180 tn 1 sn 3)wd3: status 59<rdy,seekdone,drq,err> error 40<uncorr>

followed by a hang. 

Now it's giving a panic (integer divide overflow) whenever I try to read that
block. Managed to get a dump, but kgdb won't tell me anything (can't find
kstack).

I guess I'll reformat the disks eventually; anybody got any other ideas?

Ian.