Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 2 Aug 1996 02:25:52 -0700 (PDT)
From:      "Keith Beattie[SFSU Student]" <beattie@george.lbl.gov>
To:        fqueries@jraynard.demon.co.uk (James Raynard)
Cc:        questions@freebsd.org
Subject:   Reinstall from scratch (was vnode_pager_input: I/O read error)
Message-ID:  <199608020925.CAA18883@george.lbl.gov>
In-Reply-To: <199607282009.UAA02702@jraynard.demon.co.uk> from "James Raynard" at Jul 28, 96 08:09:59 pm

next in thread | previous in thread | raw e-mail | index | archive | help
James Raynard wrote:
> 
> I would certainly try installing from scratch again.  
> 

This is kinda long, but my 2.1.0-RELEASE just crashed hard and I had
to reinstall it.  I'd like to get some input on what might have
gone wrong. TIA

<Flashback>
To refresh your memories, I had recently installed 2.1.0-RELEASE,
built a new kernel which didn't finish booting because it spun trying
to spawn getty's that kept dieing.  I had to hit reset to reboot with
the GENERIC kernel, fsck failed resulting in the loss of a few files
under /usr (which is on a 1G disk of it's own, / /var and swap are on
the 1st ide drive).  After rebooting with the GENERIC kernel I started
getting:

  vnode_pager_input: I/O read error
  vm_fault: pager input (probably hardware) error, PID 243 failure
  pid 243: cc1: uid 0: exited on signal 11

sprinkeled in with some:

  sd0(aha0:0:0): ILLEGAL REQUEST asc:24,0 Invalid field in CDB

errors.  I was going to start exploring if the sig 11's were from bad
SIMMs but then when booting, fsck never finished due to

  *** /dev/rsd0s1e
  BAD SUPERBLOCK: VALUES IN SUPERBLOCK DISAGREE WITH THOSE IN FIRST ALTERNATE
  pid 13: fsck: uid 0: exited on signal 8
  floating exception

It was then suggested that I tweak my cache BIOS settings and look for
an overheating CPU.
</flashback>

Well, I started by trying to see if DOS's himem /testmem:on would find
any problem with my RAM, but it didn't.  So, I took off the cover,
pulled out SIMM's, disabled the CPU's internal and external cache (in
the BIOS), rebooting with each change, with no luck, fsck still died
from a sig 8.  My cpu is a 486DX33 which I don't believe have a
history of overheating so the only thing left to do was reformat &
reinstall.

<sigh>

I used the windoze SCSI utilities that came with the SCSI card and
disk to reformat the drive and used M$'s scandisk to check it for any
errors: none found.  Btw, DOS and Windoze (on my 2nd ide disk) never
blinked while all these problems were happening.  Granted they don't
use the SCSI disk and they don't generally push hardware the same way
a unix system will, but it seems worth noting.

So I re-installed and compiled a new kernel.  Now things seem to be
working fine.  I'd be exagerating though, if I said I felt confident
with the system.

I'm genuinely stumped as to what happened.  I'm open to
suggestions/opinions/whatever.  Here's what's different about my
system now than before:

1) I changed a few settings in my BIOS.  (I found the documentation!)
the changed setting are: "AT BUS Clock Selection" from CLKIN/6 to
CLKIN/4 and the "Memory Read Wait State from 1 to 0".  These are the
suggested states for the 33MHz CPU.  The rest were left at the default
settings.

2) I left out a few config setting that I included on the kernel that
seemed to bring my machine down.  They are the "AUTO_EOI_2",
"DUMMY_NOPS" and "TUNE_1542" options.  LINT suggests that they might
not work for all hardware.

So, would one of, or a combination of these things bring a machine to
it's knees?  Or should I just write this off as "cosmic rays"?

Oh yea, a few short questions:

Are there any FreeBSD utilities to check for bad RAM?  The
documentation for DOS's himem.exe claims to be better than the
checking that most motherboards do on startup.  Anything like that for
FreeBSD?

Is there any documentation about BIOS settings specifically for
FreeBSD?  There are several other BIOS settings that might be tweaked
for FreeBSD, any authoritive words on this?


Tons o'thanks,
Keith

-- 
// Keith Beattie                  Lawrence Berkeley Laboratory (LBL) \\
// SFSU Grad Student   Imaging and Distributed Computing Group (ITG) \\
// KSBeattie@lbl.gov                 http://www-itg.lbl.gov/~beattie \\
// 1 Cyclotron Rd.  MS: 50B-2239  Berkeley, CA 94720  (510) 486-6692 \\



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199608020925.CAA18883>