major installation problem (sync vs async?)

major installation problem (sync vs async?)

Post by daddy2time » Fri, 31 Aug 2001 02:42:51



I have an AST Ascentia 900n laptop (circa 1994) with 486dx75mhz, 24 MB ram.
The original disk crashed, so I put in a Toshiba 6GB.  I've run linux,
win9x, FreeBSD on it before, no problem.  Now, I'm trying to put OpenBSD on
it.  I've put in about 20 hours so far with no luck (too much time invested
now to stop :)

Here's what I've done:

Boot floppy29, pcmcia (and all other hdw) recognized just fine.  I own the
2.9 CD but since the machine does not have a cdrom, I figured I'd just do
ftp install.  I tried from many sites, but once the machine has received a
couple of MB of data, it just hangs, or I get a few soft errors
(timeouts) -- the first 3 or so get corrected then it hangs.  At first, I
thought maybe it was the nic (Netgear fa410tx), it seemed to get awfully hot
(just a barely below 'too hot to touch').  I tried http install, same
problem.  I tried to create an mdos partition to hold all the install files
(*.tgz) copied from the cd via a Microsolutions Backpack, but had trouble
getting OpenBSD install to mount that partition -- then when I finally did,
it hung and wiped out the whole disk.

By the way, I'm using the whole disk for OpenBSD -- no dual booting for me!

OK, so now I'm getting pissed.  I mount the OpenBSD CD in another OpenBSD
machine, turn on ftpd, go through the install sequence, choose ftp install
from the other OpenBSD machine, and the same thing happens.  Shit!  Maybe it
is the nic overheating.  I go through the install again, and before choosing
the install method, I escape to shell, mount my home partition (wd0e), ftp
to the other OpenBSD machine, and get the smaller files; etc29.tgz,
game29.tgz, misc29.tgz, xshare29.tgz.  No problem.  Then I try man29.tgz.  I
get a couple of soft errors, but they get corrected and I get the whole
file.  Same for xbase29.tgz.  Now, all I need is base29.tgz, comp29.tgz,
xfont29.tgz, and xserv29.tgz.  I start with base29.tgz.  More soft errors
(corrected), then, HANG!  Shit!

I leave it for the night, get into bed and do a little light reading before
sleep: OpenBSD FAQ!  The next morning I give a * look to the laptop, and
go to work.  I read more FAQ on the train.  On the way home, I read more
FAQ.  I read, FAQ 11.5, Why aren't we using async mounts?.  Hmmmm!  When the
install script mounts my partitions, it says, 'blah blah blah ASYNCHRONOUS'!
More Hmmmm!

Just to see if it makes a difference, I go through the install again, escape
to shell, and try to ftp the install packages from my other OpenBSD machine,
but, this time, I've mounted my home partition with:

mount_ffs -o sync /dev/wd0e /mnt

NOW, I can get everything!  It takes a little longer, but I even get the
four big tar files.  Cool!

Now I try local disk install.  Nope.  Shit!  Again.  All four large packages
(over 10MB each) have invalid headers.

Now what?
Any ideas?
Why is install mounting all my partitions *asynchronously*?
Does that matter?
The only other thing I can add, is that after using disklabel, I get a few
sectors that cannot be used: I assume this is because the disk won't divide
nicely into whole cylinders.  (by the way, once at the shell, how do I get
back into disklabel?  Can I?  I've tried, 'disklabel -e wd0' but from there,
'p' just gives me '4.2BSD filesystem (or '512 4096 16' except on alpha, sun4
amiga, sun3...)'.  I wanted to try to print it to a file to include here,
but no go.
Below, is my dmesg.  I will gleefully supply any other info necessary.

Thanks.

OpenBSD 2.9 (RAMDISK) #295: Sat Apr 28 14:17:53 MDT 2001

cpu0: Intel 486DX4 ("GenuineIntel" 486-class)
cpu0: FPU,V86
real mem  = 24756224 (24176K)
avail mem = 19492864 (19036K)
using 327 buffers containing 1339392 bytes (1308K) of memory
mainbus0 (root)
bios0 at mainbus0: AT/286+(ee) BIOS, date 06/09/93
apm0 at bios0: Power Management spec V1.1
pcibios at bios0 function 0x1a not configured
bios0: ROM list: 0xc0000/0xa000 0xca000/0x2000
isa0 at mainbus0
isadma0 at isa0
pckbc0 at isa0 port 0x60/5
pckbd0 at pckbc0 (kbd slot)
pckbc0: using irq 1 for kbd slot
wskbd0 at pckbd0: console keyboard
vga0 at isa0 port 0x3b0/48 iomem 0xa0000/131072
wsdisplay0 at vga0: console (80x25, vt100 emulation), using wskbd0
wdc0 at isa0 port 0x1f0/8 irq 14
wd0 at wdc0 channel 0 drive 0: <TOSHIBA MK6412MAT>
wd0: 16-sector PIO, LBA, 6194MB, 13424 cyl, 15 head, 63 sec, 12685680
sectors
wd0(wdc0:0:0): using PIO mode 4
npx0 at isa0 port 0xf0/16: using exception 16
fdc0 at isa0 port 0x3f0/6 irq 6 drq 2
fd0 at fdc0 drive 0: 1.44MB 80 cyl, 2 head, 18 sec
pcic0 at isa0 port 0x3e0/2 iomem 0xd0000/16384
pcic0 controller 0: <Intel 82365SL rev 2> has socket A only
pcmcia0 at pcic0 controller 0 socket 0
pcic0: irq 3, polling enabled
pcic2 at isa0 port 0x3e4/2 iomem 0xd4000/16384
pcic2 controller 1: <Intel 82365SL rev 2> has socket A only
pcmcia1 at pcic2 controller 1 socket 0
ne3 at pcmcia1 function 0 "NETGEAR, FA410TX, Fast Ethernet" port 0x340/32
irq 4
ne3: address 00:e0:98:73:45:4b
ukphy0 at ne3 phy 0: Generic IEEE 802.3u media interface
ukphy0: OUI 0x1e0400, model 0x0000, rev. 3
pcic2: irq 9, polling enabled
biomask 4040 netmask 4050 ttymask 425a
rd0: fixed, 3560 blocks
rootdev=0x1100 rrootdev=0x2f00 rawdev=0x2f02
OpenBSD 2.9 (RAMDISK) #295: Sat Apr 28 14:17:53 MDT 2001

cpu0: Intel 486DX4 ("GenuineIntel" 486-class)
cpu0: FPU,V86
real mem  = 24756224 (24176K)
avail mem = 19492864 (19036K)
using 327 buffers containing 1339392 bytes (1308K) of memory
mainbus0 (root)
bios0 at mainbus0: AT/286+(ee) BIOS, date 06/09/93
apm0 at bios0: Power Management spec V1.1
pcibios at bios0 function 0x1a not configured
bios0: ROM list: 0xc0000/0xa000 0xca000/0x2000
isa0 at mainbus0
isadma0 at isa0
pckbc0 at isa0 port 0x60/5
pckbd0 at pckbc0 (kbd slot)
pckbc0: using irq 1 for kbd slot
wskbd0 at pckbd0: console keyboard
vga0 at isa0 port 0x3b0/48 iomem 0xa0000/131072
wsdisplay0 at vga0: console (80x25, vt100 emulation), using wskbd0
wdc0 at isa0 port 0x1f0/8 irq 14
wd0 at wdc0 channel 0 drive 0: <TOSHIBA MK6412MAT>
wd0: 16-sector PIO, LBA, 6194MB, 13424 cyl, 15 head, 63 sec, 12685680
sectors
wd0(wdc0:0:0): using PIO mode 4
npx0 at isa0 port 0xf0/16: using exception 16
fdc0 at isa0 port 0x3f0/6 irq 6 drq 2
fd0 at fdc0 drive 0: 1.44MB 80 cyl, 2 head, 18 sec
pcic0 at isa0 port 0x3e0/2 iomem 0xd0000/16384
pcic0 controller 0: <Intel 82365SL rev 2> has socket A only
pcmcia0 at pcic0 controller 0 socket 0
pcic0: irq 3, polling enabled
pcic2 at isa0 port 0x3e4/2 iomem 0xd4000/16384
pcic2 controller 1: <Intel 82365SL rev 2> has socket A only
pcmcia1 at pcic2 controller 1 socket 0
ne3 at pcmcia1 function 0 "NETGEAR, FA410TX, Fast Ethernet" port 0x340/32
irq 4
ne3: address 00:e0:98:73:45:4b
ukphy0 at ne3 phy 0: Generic IEEE 802.3u media interface
ukphy0: OUI 0x1e0400, model 0x0000, rev. 3
pcic2: irq 9, polling enabled
biomask 4040 netmask 4050 ttymask 425a
rd0: fixed, 3560 blocks
rootdev=0x1100 rrootdev=0x2f00 rawdev=0x2f02

 
 
 

major installation problem (sync vs async?)

Post by daddy2time » Fri, 31 Aug 2001 11:11:34


Should I be posting this to the mailing list?  Here?  Both?

Here's an update.

Just before choosing ftp install (from my other OpenBSD machine), I
unmounted all partitions and re-mounted them '-o sync'.  Now, the install
*seemed* to go well.  No hangs, no soft errors.  I chose a timezone, then:

making all device nodes...
Installing boot block...
boot: /mnt/boot
proto: /usr/mdec/biosboot
device: /dev/rwd0c
/usr/mdec/biosboot: entry point 0
proto bootblock size 512
room for 12 filesystem blocks at 0x16f
Will load 7 blocks of size 8192 each.
Using disk geometry of 63 sectors and 255 heads.



/mnt/boot: 3 entries total
using MBR partition 3: typ 166 (0xa6) offset 63 (0x3f)

Enabling machdep.allowaperture. Read xf86(4) for more information.

Memory fault (core dumped)
Unmounting filesystems... /mnt/tmp /mnt/var /mnt/home /mnt/usr /mnt ...Done.

CONGRATULATIONS!  You have successfull installed OpenBSD!....

#

Note the Memory fault :((

reboot gives 'Using Drive: 0 Partition: 3' and hangs.

[snip]

 
 
 

major installation problem (sync vs async?)

Post by Alfred Breu » Fri, 31 Aug 2001 15:37:33


You might want to try a 2.8 installation. There have been major
changes in 2.9 with one of the results being 2.9 pretty picky on
harddrives. [i.e. 2.9 showed fsbn errors on an old maxtor hd and
refused to continue installation while 2.8 installed w/o problems.]

Also, you might want to 'disable pciide' in ukc (just a shot in the
dark -- works with Quantum Fireballs, maybe with the Toshiba hd.

A 1993 bios is pretty old. You confirmed that the 6 GB hd is
recognized and everything translated below cyl 1024 ?

Finally, you might install just a minimum 2.9. Then just cp all other
*.tgz to the notebook and untar these.

Anyway. Wiping the partition table when installing from an msdos
partition of the same box -- I've never seen this. Sorry to say,
there's something basically wrong on your side.

 
 
 

major installation problem (sync vs async?)

Post by daddy2time » Fri, 31 Aug 2001 23:46:46


 > You might want to try a 2.8 installation. There have been major
 > changes in 2.9 with one of the results being 2.9 pretty picky on
 > harddrives. [i.e. 2.9 showed fsbn errors on an old maxtor hd and
 > refused to continue installation while 2.8 installed w/o problems.]

I'll give 2.8 a try, but...

 > Also, you might want to 'disable pciide' in ukc (just a shot in the
 > dark -- works with Quantum Fireballs, maybe with the Toshiba hd.

dmesg (or rather /kern/msgbuf) says 'pcibios at bios0 function 0x1a not
configured' and there is no further mention of pci-anything, so I assume
pciide is already automatically disabled.  Plus, 'wdc0 at isa0 port 0x1f0/8
irq 14'.  Now I am wondering if maybe the machine is too slow to handle PIO
mode 4.  This is getting to be a bit over my head, but could I tell it to
drop to a lower PIO mode?  Could I even run ukc on the install kernal?

 > A 1993 bios is pretty old. You confirmed that the 6 GB hd is
 > recognized and everything translated below cyl 1024 ?

Well, as I said, I've run linux, win95 and 98 in this hardware configuration
before with no problems so I assume it is ok.

 > Finally, you might install just a minimum 2.9. Then just cp all other
 > *.tgz to the notebook and untar these.
 >
 > Anyway. Wiping the partition table when installing from an msdos
 > partition of the same box -- I've never seen this. Sorry to say,
 > there's something basically wrong on your side.

You are right, that was very strange, but I still wonder about PIO mode 4.
*AND* I'm still very curious about why FAQ 11.5 says asychronous is *not*
good, but the install mounts the partitions asynchronously.  Is that one of
the changes from 2.8 to 2.9?

Thanks

 
 
 

major installation problem (sync vs async?)

Post by daddy2time » Sat, 01 Sep 2001 13:24:53


2.8 fails miserably too.

I'm assuming it is a hardware problem, but I'd like to narrow it down if
possible.

Can anyone tell me if this group is, in any way, connected with the mailing
list?

Perhaps I should post there?

Thanks...


Quote:> > You might want to try a 2.8 installation. There have been major
>  > changes in 2.9 with one of the results being 2.9 pretty picky on
>  > harddrives. [i.e. 2.9 showed fsbn errors on an old maxtor hd and
>  > refused to continue installation while 2.8 installed w/o problems.]

> I'll give 2.8 a try, but...

>  > Also, you might want to 'disable pciide' in ukc (just a shot in the
>  > dark -- works with Quantum Fireballs, maybe with the Toshiba hd.

> dmesg (or rather /kern/msgbuf) says 'pcibios at bios0 function 0x1a not
> configured' and there is no further mention of pci-anything, so I assume
> pciide is already automatically disabled.  Plus, 'wdc0 at isa0 port
0x1f0/8
> irq 14'.  Now I am wondering if maybe the machine is too slow to handle
PIO
> mode 4.  This is getting to be a bit over my head, but could I tell it to
> drop to a lower PIO mode?  Could I even run ukc on the install kernal?

>  > A 1993 bios is pretty old. You confirmed that the 6 GB hd is
>  > recognized and everything translated below cyl 1024 ?

> Well, as I said, I've run linux, win95 and 98 in this hardware
configuration
> before with no problems so I assume it is ok.

>  > Finally, you might install just a minimum 2.9. Then just cp all other
>  > *.tgz to the notebook and untar these.

>  > Anyway. Wiping the partition table when installing from an msdos
>  > partition of the same box -- I've never seen this. Sorry to say,
>  > there's something basically wrong on your side.

> You are right, that was very strange, but I still wonder about PIO mode 4.
> *AND* I'm still very curious about why FAQ 11.5 says asychronous is *not*
> good, but the install mounts the partitions asynchronously.  Is that one
of
> the changes from 2.8 to 2.9?

> Thanks

 
 
 

1. ext2fs sync vs. async speed and safety

How much different is using sync vs. async on ext2fs in speed
and safety?

--
------------------------------------------------------------------------

Unsolicited bulk or commercial email is not welcome.             netcom.com
No warranty of any kind is provided with this message.

2. In wich order does the run time linker search the lib*so ?

3. mount ext2fs sync vs. async

4. halt reboot for users

5. swap priority, sync vs. async

6. I'm sick of UnixWare 1.1, looking at Solaris...need info

7. Linux async vs. FreeBSD sync

8. uugetty for 4.2?

9. "sync;sync" vs. "sync /important_only &"

10. Major problem with disc syncing

11. EDO vs FP ram, async vs burst-mode cache

12. Problem adding tty port using 128 Port Async Adapter with 16 port Async Node

13. Serial Ports: Sync or Async