Date: Mon, 3 Nov 1997 20:45:43 +0100 (MET) From: "Julian H. Stacey" <jhs@freebsd.org> To: hackers@freebsd.org Subject: bad144 & ide format Message-ID: <199711031945.UAA00778@flip.jhs.no_domain>
next in thread | raw e-mail | index | archive | help
Help ! :-)
I've config'd & used loads of SCSI drives on FreeBSD systems,
but I never seem to be able to correctly add IDE drives to pre-built systems ...
I have a 1 gig (on a modern ASUS mboard that supports big discs),
& 2 * 200M & a 40M on other boxes,
Sod's law dictates the only one that works trouble free with my FreeBSD boxes
is the 40M :-( (maybe 'cos its so small I rarely use it ( it just serves as
/usr/share on a sacrificial crash host )).
I'm definitely doing something wrong, question is what ?
all I want to do is format them, run a bad144 on them to fix
the bad blocks, disklabel & newfs & add them to existing systems.
No big deal, been doing this stuff for years,
even used to regularly run bad144 on my st506 on a BSD 4.2 NSC32016,
& have been swapping BSD discs around on intel & nsc BSDs for years ...
_But_ I must have a blind spot on IDE,
the bad144 is going wrong (where I marked **** in typical log below)
& I'd be grateful to anyone who spots what I'm doing wrong.
I did look at the handbook first, but to no avail.
PS I've been using 2.2.2 based systems, but have looked on
FreeBSD current source & cvs & there's no format.8 despite
man bad144 pointing to it,
only format 8 I can see is in my NetBSD current tree:
src/share/man/man8/man8.hp300/format.8
src/share/man/man8/man8.vax/format.8
PS please ensure a cc: jhs@freebsd.org as I'm only on current@,
not hackers@.
==========
Log
Under AMI Bios:
Set to IDE Type 47: Cyl=1366, Heads=8, Sec=38, 203MB.
format, interleave, media check, all with Mark_Bad_Tracks=No
Boot FreeBSD
dmesg:
wdc0: unit 0 (wd0): <Conner Peripherals 210MB - CP3201>
wd0: 203MB (415872 sectors), 684 cyls, 16 heads, 38 S/T, 512 B/S
This probe result is suprising because 1366/2=683
bad144 /dev/rwd0
Bad pack magic number (pack is unlabeled)
bad144 -v -s -c /dev/rwd0
Bad pack magic number (pack is unlabeled)
vi disktab:
c_wd0|Conner_210: \
:dt=ST506:ty=winchester:nt#16:nc#684:ns#38:su#415872: \
:pa#415872: :oa#0: :ta=4.2BSD: :fa#1024: :ba#8192: \
:pc#415872: :oc#0: :tc=unused:
disklabel -w -r wd0 c_wd0
bad144 -v -s -c /dev/rwd0
runs OK (takes a while) then come errors
bad144 /dev/rwd0
cyl: 684, tracks: 16, secs: 38, sec/cyl: 608, start: 0, end: 415708
415264 of 415708 blocks ( 99%)
bad block information at sector 415834 in /dev/rwd0:
cartridge serial number: 1391109348(10)
bt_flag=c598(16)?
sn=2148934, cn=3523, tn=180, sn=110
bad144: /dev/rwd0: bad flag in bad-sector table
bad144: /dev/rwd0: bad magic number
bad144: cyl/trk/sect out of range in existing entry: sn=2148934, cn=3523, tn=180, sn=110
bad144: cyl/trk/sect out of range in existing entry: sn=34813762, cn=57244, tn=243, sn=176
***** + lots more errors ********************************************
Create a new entry:
vi disktab:
c_wd00|Conner_210: \
:dt=ST506:ty=winchester:nt#16:nc#683:ns#38:su#415264: \
:pa#415264: :oa#0: :ta=4.2BSD: :fa#1024: :ba#8192: \
:pc#415264: :oc#0: :tc=unused:
dc
16 38 683 * * p 415264
16 38 684 * * p 415872
disklabel -w -r wd0 c_wd00
open partition would move or shrink
halt
kernel -s
open partition would move or shrink
Need to zap the label, so try with more cylinders, maybe CT mag was wrong,
& FreeBSD probe is correct in suggesting 2 more cylinders of 8 heads available
Bios:
Set to Type 47: Cyl=1368, Heads=8, Sec=38, 203MB
format & interleave Mark_Bad_Tracks=Yes; media analysis
dmesg:
<Conner Peripherals 210MB - CP3201>
203MB (415872 sectors), 684 cyls, 16 heads, 38 S/T, 512 B/S
disklabel -w -r wd0 c_wd0
chris /kernel: wd0: invalid primary partition table: no magic
bad144 -v -s -c /dev/rwd0
415264 of 415708 blocks ( 99%)
bad block information at sector 415834 in /dev/rwd0:
cartridge serial number: 1840700269(10)
bt_flag=db6d(16)?
sn=17083423, cn=28086, tn=182, sn=219
bad144: /dev/rwd0: bad flag in bad-sector table
bad144: /dev/rwd0: bad magic number
bad144: cyl/trk/sect out of range in existing entry: sn=17083423, cn=28086, tn=182, sn=219
bad144: cyl/trk/sect out of range in existing entry: sn=34157508, cn=56173, tn=109, sn=182
bad144: cyl/trk/sect out of range in existing entry: sn=28469519, cn=46811, tn=219, sn=109
..... + lots more errors .....
==========
PPS I've not tried formatting IDEs on the ASUS board, 'cos it's not an AMI
BIOS & has no support for IDE format in EPROM, so Ive been using the older
pre LBA era AMI BIOSes on the Gigabyte & another non ASUS old mboard,
but although that could explain other LBA problems with the 1Gig drive,
there's no reason the 200M IDEs
shouldn't just format & bad144 cleanly ... if only I knew the trick ?
The strange thing is, despite bad144 producing these error messages,
I can newfs a 200M file system & write & read check data to the FS
without consolke errors, but I must be walking a tightrope, in light of the
bad144 error report.
Julian
--
Julian H. Stacey
Home <jhs@freebsd.org> http://www.freebsd.org/~jhs/ Tel. +49 89 268616
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199711031945.UAA00778>