From owner-freebsd-questions  Tue Jun 13 14:21:58 1995
Return-Path: questions-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id OAA15493
          for questions-outgoing; Tue, 13 Jun 1995 14:21:58 -0700
Received: from iaehv.IAEhv.nl (root@iaehv.IAEhv.nl [192.87.208.2])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id OAA15479
          for <questions@freebsd.org>; Tue, 13 Jun 1995 14:21:55 -0700
Received: by iaehv.IAEhv.nl (8.6.12/1.63)
    id XAA08095; Tue, 13 Jun 1995 23:21:09 +0200
From: wjw@IAEhv.nl (Willem Jan Withagen)
Message-Id: <199506132121.XAA08095@iaehv.IAEhv.nl>
X-Disclaimer: iaehv.nl is a public access UNIX system and cannot be held
      	      responsible for the opinions of its individual users.
Subject: Re: Medium error
To: dufault@hda.com (Peter Dufault)
Date: Tue, 13 Jun 1995 23:21:08 +0200 (MET DST)
Cc: wjw@IAEhv.nl, questions@freebsd.org
In-Reply-To: <199506132108.RAA13083@hda.com> from "Peter Dufault" at Jun 13, 95 05:08:29 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 2239      
Sender: questions-owner@freebsd.org
Precedence: bulk

You ( Peter Dufault ) write:
=>  
=>  Willem Jan Withagen writes:
=>  
=>  (Re-directed to -questions)
=>  
=>  > We're running an ISP site and our newsdisk is giving us a hard time, but
=>  > it's real hard to replace the damn thing at the moment.
=>  > 
=>  > What do we have:
=>  > 	A P90-asus with NCR controller with a quantum 2.1 Empire as sd2.
=>  
=>  Be sure you provide all the information about your installation.  In
=>  this case, I need to know which release you are running.

Eh, dumb, dumb, dumb. I keep telling this to our users over and over.
Boy do I feel stuppid:
	Running 2.0R with a few minor patches 

./i386/conf/files.i386.orig
./i386/pci/ncr.c.orig
./i386/pci/ncrreg.h.orig
./kern/kern_sysctl.c.orig
./kern/vfs_syscalls.c.orig
./netinet/ip_input.c.orig
./netinet/ip_output.c.orig
./netinet/ip_fw.c.orig
./netinet/ip_fw.h.orig
./netinet/raw_ip.c.orig
./netinet/in.h.orig
./sys/malloc.h.orig

These we're added by our local FreeBSD guru, the ncr patches even after
consults with the NCR author.

=>  > And this baby is generating a medium error on just one location. 
=>  > Jun 13 22:19:50 iaehv kernel: sd2(ncr0:2:0): medium error, info = 2817606 (decimal)
=>  
=>  Is this always the same "info" code?

Yes, this is the only error spot it reports always this 2817606

=>  > The funny (??) thing is that this cause I/O errors in INN in serveral spool
=>  > directories. Even after I created a real large file to try and cover the bad
=>  > spot with just-another-file.
=>  > 
=>  > So I'm starting to wonder: can the bad spot contain more than just one
=>  > directory which is getting trouble?
=>  
=>  It is probably a single spot.

That's what I hoped. But it still doesn't give me any hint as to why I now
can not create new files in 6 different directories?
And doing a 'touch a' in such a directory immediatly triggers the kernel
message.

=>  > But the real serious question: Is there any way of getting this ONE error
=>  > out of the way by just any reasonable method?
=>  
=>  I'll make some suggestions after I see what you're running.  You have
=>  more options if you are running 2.05.

That's what Guido told me, but we'd want to test-run 2.05 before letting it
take a whack at our users. :-)

--WjW