RAID controllers - BIOS conflict - pls help

[nobody]

"I have a MSI Master2-FAR (VIA K8T800 chipset) dual Opteron 242 system
with boot drive on built-in VIA SATA RAID controller. I installed
Promise FastTrak SX4100 (SATA RAID controller - 32 bit PCI, hardware
RAID 5 engine) and 3 WD 250GB drives. When I configured the drives as
RAID 5, I received the message "PCI ROM space is NOT enough", and then
the board was unable to load VIA RAID controller to boot off it.
However, when I connected the same drive to a free port on SX4100
instead of onboard SATA, the system booted just fine (Win2k), and both
RAID controllers were in device manager. However this solution is not
acceptable because my intent was to have bootable RAID1 on VIA and
RAID5 on SX4100 reserved for data. Please let me know how to ensure
that Promise SX4100 BIOS loads AFTER VIA SATA BIOS, so that my old
boot configuration could be restored."

The msg above is a copy of what I sent to both MSI and Promise
support. So far no reply from any of them.

Google/Yahoo have a couple of hits on "PCI ROM space is NOT enough",
but only complaints and no solutions. Seems like Promise grabs so
much memory out of whatever is allocated for BIOS that no other device
can be loaded after that. Funny though that as long as no logical
drive is defined, everything goes fine and VIA SATA is recognized on
boot. Could not find anything in either SX4100 setup or mobo BIOS
setup that would help. Could it be that VIA SATA sucks? Would
another Promise/sil RAID controller solve the issue? All I want is
RAID1 as system drive and RAID5 as data drive. I guess I could go
SCSI for boot drive in worst case, but there is no guarantee that SCSI
BIOS would not conflict with Promise BIOS, either.

Any and all help will be greatly appreciated.

TIA

NNN

 

[George Macdonald]

"I have a MSI Master2-FAR (VIA K8T800 chipset) dual Opteron 242 system
with boot drive on built-in VIA SATA RAID controller. I installed
Promise FastTrak SX4100 (SATA RAID controller - 32 bit PCI, hardware
RAID 5 engine) and 3 WD 250GB drives. When I configured the drives as
RAID 5, I received the message "PCI ROM space is NOT enough", and then
the board was unable to load VIA RAID controller to boot off it.
However, when I connected the same drive to a free port on SX4100
instead of onboard SATA, the system booted just fine (Win2k), and both
RAID controllers were in device manager. However this solution is not
acceptable because my intent was to have bootable RAID1 on VIA and
RAID5 on SX4100 reserved for data. Please let me know how to ensure
that Promise SX4100 BIOS loads AFTER VIA SATA BIOS, so that my old
boot configuration could be restored."

The msg above is a copy of what I sent to both MSI and Promise
support. So far no reply from any of them.

Google/Yahoo have a couple of hits on "PCI ROM space is NOT enough",
but only complaints and no solutions. Seems like Promise grabs so
much memory out of whatever is allocated for BIOS that no other device
can be loaded after that. Funny though that as long as no logical
drive is defined, everything goes fine and VIA SATA is recognized on
boot. Could not find anything in either SX4100 setup or mobo BIOS
setup that would help. Could it be that VIA SATA sucks? Would
another Promise/sil RAID controller solve the issue? All I want is
RAID1 as system drive and RAID5 as data drive. I guess I could go
SCSI for boot drive in worst case, but there is no guarantee that SCSI
BIOS would not conflict with Promise BIOS, either.

Any and all help will be greatly appreciated.

Dunno about VIA SATA sucking but I'm pretty sure this is a MSI BIOS
issue - a check at MSI Forum shows that people are getting similar ("Not
enough space to copy PCI option ROM 00:13:00") msgs even with a mix of
on-board chipset (VIA & nVidia) SATA RAID and on-board SIL SATA RAID.
Obviously there is a conflict between address spaces for the two RAID
firmwares in the Reserved Memory space (640K -> 1M)... if the msg is to be
believed.

If you're booting off the VIA RAID, is it not possible to just disable the
Promise BIOS and let the OS driver load handlle the SX4100 RAID? Are you
sure the SX4100 BIOS is not set up to boot from the RAID-5?... thus causing
a conflict in boot device? ISTR that with SCSI, you can choose to not hook
the SCSI BIOS if there's no need to boot of a SCSI device.

 

[nobody]

On Mon, 15 May 2006 05:58:22 -0400, George Macdonald

....snip...

Dunno about VIA SATA sucking but I'm pretty sure this is a MSI BIOS
issue - a check at MSI Forum shows that people are getting similar ("Not
enough space to copy PCI option ROM 00:13:00") msgs even with a mix of
on-board chipset (VIA & nVidia) SATA RAID and on-board SIL SATA RAID.
Obviously there is a conflict between address spaces for the two RAID
firmwares in the Reserved Memory space (640K -> 1M)... if the msg is to be
believed.

Got response from MSI that flashing the latest BIOS would fix it. But
it is marked as beta, and I am hesitant to use any beta software, let
alone flashing it. Also have a problem finding in my pile of spare
parts a floppy drive (don't even have it installed in this system),
and MSI site explicitly warns against using any other boot device to
flash ;-)

If you're booting off the VIA RAID, is it not possible to just disable the
Promise BIOS and let the OS driver load handlle the SX4100 RAID? Are you

If I hit <Esc> when prompted on boot, Promise BIOS gets skipped, VIA
SATA loads normally, but then Promise thinks the drives are not
assigned to any logical drive, and suggest to build it. It builds
RAID5, lets it be formatted, but then after the next reboot thinks
there is no logical drive defined, and suggests to rebuild the RAID
again... and again...and again... Maybe it's different for other RAID
levels, but looks that 5 must be loaded in BIOS regardless of being
bootable.

sure the SX4100 BIOS is not set up to boot from the RAID-5?... thus causing
a conflict in boot device? ISTR that with SCSI, you can choose to not hook
the SCSI BIOS if there's no need to boot of a SCSI device.

Couldn't find anything close to that, in either Promise BIOS setup or
their GUI-based tool. Will wait for Promise reply. If they don't
suggest anything short of flashing the mobo, I'll have to take a deep
breath and flash that beta BIOS...

But thanks anyway

NNN

 

[George Macdonald]

On Mon, 15 May 2006 05:58:22 -0400, George Macdonald

...snip...
Got response from MSI that flashing the latest BIOS would fix it. But
it is marked as beta, and I am hesitant to use any beta software, let
alone flashing it. Also have a problem finding in my pile of spare
parts a floppy drive (don't even have it installed in this system),
and MSI site explicitly warns against using any other boot device to
flash ;-)

Huh? I haven't seen that in my MSI adventures - in fact, although their
instructions say to boot from floppy, they say *not* to run Flash from the
floppy and recommend strongly to run the flash from a hard drive (FAT,
FAT32) or from a Win98 Startup RAMDRIVE (NTFS). You should be OK booting
from a USB flash drive, though I've never tried it. See here
http://h18000.www1.hp.com/support/files/serveroptions/us/download/20306.html
for a utility to create a bootable USB flash drive, though it needs a
source for the DOS boot files.

 

[nobody]

Huh? I haven't seen that in my MSI adventures - in fact, although their
instructions say to boot from floppy, they say *not* to run Flash from the
floppy and recommend strongly to run the flash from a hard drive (FAT,
FAT32) or from a Win98 Startup RAMDRIVE (NTFS). You should be OK booting
from a USB flash drive, though I've never tried it. See here
http://h18000.www1.hp.com/support/files/serveroptions/us/download/20306.html
for a utility to create a bootable USB flash drive, though it needs a
source for the DOS boot files.

Maybe this is an aberration of the memory. Even if they don't say
this now, they used to. Or maybe some other place - like the mobo
manual, not even necessarily the manual for this mobo. I didn't
invent the idea that floppy is a must for flashing BIOS, honestly.
But thanks for the idea to boot from USB - this may be an easy way
out, though not without risk - as in the thread you started recently.
Flashing from a corrupt USB drive may be fatal. However I've seen a
lot more dude floppies than flash drives ;-)
Promise still silent (I don't take for an answer the autoreply that
they received my request) - that much for customer service. Looks
like I am doomed to try and flash that beta BIOS.

NNN

 

[nobody]

Maybe this is an aberration of the memory. Even if they don't say
this now, they used to. Or maybe some other place - like the mobo
manual, not even necessarily the manual for this mobo. I didn't
invent the idea that floppy is a must for flashing BIOS, honestly.
But thanks for the idea to boot from USB - this may be an easy way
out, though not without risk - as in the thread you started recently.
Flashing from a corrupt USB drive may be fatal. However I've seen a
lot more dude floppies than flash drives ;-)
Promise still silent (I don't take for an answer the autoreply that
they received my request) - that much for customer service. Looks
like I am doomed to try and flash that beta BIOS.

NNN

Flashed 1.35 (beta) - didn't fix the issue, but win2k would bluescreen
just before inviting to ctrl-alt-del to log on. Would boot only safe
mode. Reflashing 1.1 fixed the Windows, but still have issues with
RAIDs.

As a side note: an old IDE HDD (under 1GB - don't even understand why
I forgot to throw it away) that already had old win95 system on it was
perfect as a flashing device - I only had to hit F8 and pick Safe mode
- Command prompt option to prevent win95 from trying to start GUI.

Looks like this is my last MSI board. :-( Will be shopping Asus or
Tyan for the next one.

NNN

 

[George Macdonald]

Flashed 1.35 (beta) - didn't fix the issue, but win2k would bluescreen
just before inviting to ctrl-alt-del to log on. Would boot only safe
mode. Reflashing 1.1 fixed the Windows, but still have issues with
RAIDs.

Just another thought: did you try turning off BIOS shadowing/caching to see
if that affects the conflict with Promise? Of course one would *think*
that PnP would be mature enough now to have fixed issues like this... and
they've taken away the jumpers which allowed manually choosing address
mapping umm, because PnP is supposed to err, "work". Lose on the
swings.... lose on the roundabouts.<shrug>

Another thought: the BIOS in my MSI K8N Neo2 Platinum seems to have a hook
at d000:0 which hints at being for USB. On the off-chance that your mbrd
is done by the same BIOS writers and that the USB hook is for booting from
a USB drive, you might try turning off USB support completely to see if
that leaves a slot for Promise to take.

As a side note: an old IDE HDD (under 1GB - don't even understand why
I forgot to throw it away) that already had old win95 system on it was
perfect as a flashing device - I only had to hit F8 and pick Safe mode
- Command prompt option to prevent win95 from trying to start GUI.

Looks like this is my last MSI board. :-( Will be shopping Asus or
Tyan for the next one.

MSI is usually pretty good with BIOS updates for their desktop boards...
but it looks like their committment to server is a bit weak - that board
has been around long enough now to be fixed.

You might want to check willbill's thread (e-mail address removed) with
his experience of SuperMicro - they look like they might be getting close
to coming out of the closet on their AMD mbrds.