Discussion:
Adaptec SCSI card with no device errors question
(too old to reply)
Mike Luther
2007-11-14 03:20:58 UTC
Permalink
I can't recall where I noticed this but want to follow up on it.

Somewhere I saw where it was posted that having an Adaptec SCSI device driver
card in an OS/2 box which did not connect with anything caused errors in the
OS/2 operations.

Can anyone re-visit or furnish data on this? Reason being is that I'd like to
create a generic OS/2 box for mobile drive tray debugging, backup and cloning
use that could handle both SATA drive trays and SCSI drive trays at the same
time. If I were to plug in a SATA drive tray with the appropriate drive but
not a SCSI tray, would this result in what was claimed?

As well, what if I were using a SCSI drive tray for the actual hard drive, but
no SATA drive tray in that slot or slots?

Thanks!
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Steve Wendt
2007-11-14 03:29:01 UTC
Permalink
Post by Mike Luther
Somewhere I saw where it was posted that having an Adaptec SCSI device
driver card in an OS/2 box which did not connect with anything caused
errors in the OS/2 operations.
I'm pretty sure the OS/2 boot CD loads the Adaptec drivers, so I believe
that would make this rumor false. Similarly, the IDE drivers (IBM1S506
and DANIS506) don't hurt anything when not needed.

If you find exceptions to this, be sure to let us know.
Michael Lueck
2007-11-14 13:55:53 UTC
Permalink
Post by Steve Wendt
Post by Mike Luther
Somewhere I saw where it was posted that having an Adaptec SCSI device
driver card in an OS/2 box which did not connect with anything caused
errors in the OS/2 operations.
I'm pretty sure the OS/2 boot CD loads the Adaptec drivers, so I believe
that would make this rumor false. Similarly, the IDE drivers (IBM1S506
and DANIS506) don't hurt anything when not needed.
I think Mike has the situation that he has the card in the computer, just that nothing is attached to the card.

A different situation from the OS/2 boot process quietly failing to load drivers if no compatible hardware is detected.
--
Michael Lueck
Lueck Data Systems
http://www.lueckdatasystems.com/
Mike Luther
2007-11-15 03:35:54 UTC
Permalink
Hi Mike!
Post by Michael Lueck
I think Mike has the situation that he has the card in the computer,
just that nothing is attached to the card.
A different situation from the OS/2 boot process quietly failing to load
drivers if no compatible hardware is detected.
Spot on correct. What I want is a single modern level motherboard and CPU box
in which I can jam an SATA drive, a SCSI drive or even an IDE drive in the
appropriate drive tray and go, go, go!

Properly set up in the CONFIG.SYS for such things, of course, grin.
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Mike Luther
2007-11-15 03:34:31 UTC
Permalink
Hi Mike!
Post by Michael Lueck
I think Mike has the situation that he has the card in the computer,
just that nothing is attached to the card.
A different situation from the OS/2 boot process quietly failing to load
drivers if no compatible hardware is detected.
Spot on correct. What I want is a single modern level motherboard and CPU box
in which I can jam an SATA drive, a SCSI drive or even an IDE drive in the
appropriate drive tray and go, go, go!

Properly set up in the CONFIG.SYS for such things, of course, grin.
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Scott G.
2007-11-14 21:31:27 UTC
Permalink
Post by Mike Luther
I can't recall where I noticed this but want to follow up on it.
Somewhere I saw where it was posted that having an Adaptec SCSI device
driver card in an OS/2 box which did not connect with anything caused
errors in the OS/2 operations.
Can anyone re-visit or furnish data on this? Reason being is that I'd
like to create a generic OS/2 box for mobile drive tray debugging,
backup and cloning use that could handle both SATA drive trays and SCSI
drive trays at the same time. If I were to plug in a SATA drive tray
with the appropriate drive but not a SCSI tray, would this result in
what was claimed?
As well, what if I were using a SCSI drive tray for the actual hard
drive, but no SATA drive tray in that slot or slots?
It will make no difference unless you're crazy enough to load the
aha154x driver. Then it's not a matter of driver letters as much as
toasting the entire memory management system.
Mike Luther
2007-11-15 03:30:54 UTC
Permalink
Hi Scott ..
Post by Scott G.
Post by Mike Luther
I can't recall where I noticed this but want to follow up on it.
Somewhere I saw where it was posted that having an Adaptec SCSI device
driver card in an OS/2 box which did not connect with anything caused
errors in the OS/2 operations.
Can anyone re-visit or furnish data on this? Reason being is that I'd
like to create a generic OS/2 box for mobile drive tray debugging,
backup and cloning use that could handle both SATA drive trays and
SCSI drive trays at the same time. If I were to plug in a SATA drive
tray with the appropriate drive but not a SCSI tray, would this result
in what was claimed?
As well, what if I were using a SCSI drive tray for the actual hard
drive, but no SATA drive tray in that slot or slots?
It will make no difference unless you're crazy enough to load the
aha154x driver. Then it's not a matter of driver letters as much as
toasting the entire memory management system.
Thank you for your help Scott. You're still a treasure for all of us here. I
have *NO* intention of ever visiting the aha154x driver ever again in all of
this after watching all the development furor you and the folks went through
over it!

As well, if I do try this in a single box, I think with the Intel 915GAVL board
planned for it, that I can force-coordinate the IRQ assignments to the PCI slot
card work such that it is fixed over what they are like for the drive tray use
no matter what that use is. If I can carefully work that out with the on-board
internal IRQ use which will keep me away from mis-adventures with audio support
and so on, having a box with whatever is needed with whatever hard drive is
jammed into the SCSI or SATA drive carrier - or even the IDE device - would
sure help in cutting down on equipment pile up, service time and research of
all kinds.

Thanks again. I hope everything is well for you and the people to which you
are close.
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Steve Wendt
2007-11-15 03:39:09 UTC
Permalink
Post by Mike Luther
As well, if I do try this in a single box, I think with the Intel
915GAVL board planned for it
I hope you are planning to use a 29160, not a 2940. The 2940 does not
function correctly with the faster PCI bus.
Mike Luther
2007-11-15 13:41:33 UTC
Permalink
Blessings to you Steve!
Post by Steve Wendt
Post by Mike Luther
As well, if I do try this in a single box, I think with the Intel
915GAVL board planned for it
I hope you are planning to use a 29160, not a 2940. The 2940 does not
function correctly with the faster PCI bus.
I think you've just answered a glitch I've seen. My workhorse Intel 915GAVL
box for backup storage in an Integrand Research rack case with SCSI DAT tape
and CD-ROM burner, has a 29160 system with an external rack unit containment
unit. It works perfectly with that 29160. That includes, as far as I've been
able to use and test, solid SCSI DAT tape backup and collection for jam-in SCSI
hard drive mobile tray operations as well as DFSEE clone work.

A second external 915GAVL box with a 2940U2W Pro seems to have long SCSI block
write issues on a fairly frequent basis with the SCSI DAT tape drive in it. I
don't see these errors unless the SCSI hard disk involved is pretty large in
this particular box. I bet you just answered the reason for that! Though
expensive, I'll order another 29160 card today for this case. That with
modifying the OS/2 Utility Floppy diskette for the 29160 vs. the 2940 code.

Then, I'll contemplate moving all the rest of the SCSI boxes to the 29160 cards
to plan toward the future, based on what you've just posted as well as what
I've, as posted, seen. That will also help standardize the OS/2 Utility Disk
floppy operations from individual sets per box to one set for multiple boxes in
the future.

I've had no trouble with the fixed MCP2 Utility diskette creation operation and
the 29160 code either with the major relay rack case, if you follow the
creation fixup disk space option Scott and crew got us with the later MCP2
release(s). From which, for all these Utility Disk sets, in general, I'd also
removed the old 1540 driver, snoop code and vectors in the utility floppy
diskettes I've got for all these newer systems already. That before I even
posted all this, from Scott's read-me information on Kernel upgrades.

Sorry for all the words here Steve but just trying to help people as well as
learn from all you folks with more awareness than I've got!
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Mike Luther
2007-11-20 23:15:01 UTC
Permalink
Question Steve ..
Post by Steve Wendt
Post by Mike Luther
As well, if I do try this in a single box, I think with the Intel
915GAVL board planned for it
I hope you are planning to use a 29160, not a 2940. The 2940 does not
function correctly with the faster PCI bus.
Is this a result of 160/320 speed SCSI drives with the 2940 cards, and not with
the older Ultrawide type SCSI drives?

Also, there are I think three basic models of 2940 Adaptec cards. The 2940UW,
the 2940UWPro and per my recall the 2940U2W card. Do we see these errors on,
for example the 320 speed current level SCSI drives against the 2940U2W with
the separated chipset for the two 68 pin channels?

I'm assuming here that we're talking about all LW type Seagate drive hardware,
the 68-pin cable and connector hardware. As well I do realize that if you are
going to use the 160/320 speed level drives you also must be prepared to use
the high quality high speed drive cables to work with them, even at 68-pin levels.

Thanks Steve ..
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Steve Wendt
2007-11-21 03:45:40 UTC
Permalink
Post by Mike Luther
Post by Steve Wendt
I hope you are planning to use a 29160, not a 2940. The 2940 does not
function correctly with the faster PCI bus.
Is this a result of 160/320 speed SCSI drives with the 2940 cards, and
not with the older Ultrawide type SCSI drives?
The drives aren't the problem, it's the PCI card that can't handle it.
Post by Mike Luther
Also, there are I think three basic models of 2940 Adaptec cards. The
2940UW, the 2940UWPro and per my recall the 2940U2W card.
I'm pretty sure the variant of 2940 is irrelevant.
Mike Luther
2007-11-21 14:49:42 UTC
Permalink
(o).(o)< .. scowl ..
The drives aren't the problem, it's the PCI card that can't handle it.
I'm pretty sure the variant of 2940 is irrelevant.
Well there goes a quick few US$ grand here and there, sigh. But I sure thank
you for the very important information.
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Trevor Hemsley
2007-11-21 19:12:13 UTC
Permalink
On Wed, 21 Nov 2007 14:49:42 UTC in comp.os.os2.bugs, Mike Luther
Post by Mike Luther
(o).(o)< .. scowl ..
The drives aren't the problem, it's the PCI card that can't handle it.
I'm pretty sure the variant of 2940 is irrelevant.
Well there goes a quick few US$ grand here and there, sigh. But I sure thank
you for the very important information.
The U2W variant won't be affected by the problem I suspect as it's a few years
newer and is LVD enabled. I'm also unaware of any problems with the ordinary
2940 series of cards other than the fact that they only run at 33MHz.
--
Trevor Hemsley, Brighton, UK
Trevor dot Hemsley at ntlworld dot com
Steven Levine
2007-11-21 18:43:02 UTC
Permalink
In <***@trevor2.dsl.pipex.com>, on 11/21/2007
at 01:12 PM, "Trevor Hemsley" <***@mytrousers.ntlworld.com>
said:

Hi,
Post by Trevor Hemsley
The U2W variant won't be affected by the problem I suspect as it's a few
years newer and is LVD enabled. I'm also unaware of any problems with
the ordinary 2940 series of cards other than the fact that they only run
at 33MHz.
FWIW, I use a leftover 2940UW to attach the scanner on a 3GHz P4 box and
it's never caused any problems.

Even if they were stable, I would not recommend using 2940UW's on 160/320
drives because it would be a waste of the drive's capabilities.

Steven
--
--------------------------------------------------------------------------------------------
Steven Levine <***@earthlink.bogus.net> MR2/ICE 3.00 beta 09pre #10183
eCS/Warp/DIY/14.103a_W4 www.scoug.com irc.ca.webbnet.info #scoug (Wed 7pm PST)
--------------------------------------------------------------------------------------------
Steve Wendt
2007-11-22 07:41:41 UTC
Permalink
Post by Steven Levine
FWIW, I use a leftover 2940UW to attach the scanner on a 3GHz P4 box and
it's never caused any problems.
Maybe for slower, non-disk devices, there are no issues. I only ever
tried hanging disk drives off mine.
Steven Levine
2007-11-23 07:20:24 UTC
Permalink
In <Voa1j.2735$***@newssvr22.news.prodigy.net>, on 11/21/2007
at 11:41 PM, Steve Wendt <***@forgetit.org> said:

Hi,
Post by Steve Wendt
Maybe for slower, non-disk devices, there are no issues. I only ever
tried hanging disk drives off mine.
I never tried this. When I switched to U160 drives, it made sense to
install a 29160 to take advantage of the drive speed.


Steven
--
--------------------------------------------------------------------------------------------
Steven Levine <***@earthlink.bogus.net> MR2/ICE 3.00 beta 09pre #10183
eCS/Warp/DIY/14.103a_W4 www.scoug.com irc.ca.webbnet.info #scoug (Wed 7pm PST)
--------------------------------------------------------------------------------------------
Steve Wendt
2007-11-22 07:40:27 UTC
Permalink
Post by Trevor Hemsley
I'm also unaware of any problems with the ordinary
2940 series of cards other than the fact that they only run at 33MHz.
I tried using my 2940UW (which works very well on my Pentium II system)
on the same motherboard Mike referred to (i915GAV), and quickly had
problems. I can't speak for the U2W.
Mike Luther
2007-11-24 02:19:27 UTC
Permalink
Aha .. I think this is focusing..
Post by Steve Wendt
I'm also unaware of any problems with the ordinary 2940 series of
cards other than the fact that they only run at 33MHz.
I tried using my 2940UW (which works very well on my Pentium II system)
on the same motherboard Mike referred to (i915GAV), and quickly had
problems. I can't speak for the U2W.
Exactly the same experience here Steve. But what I did to 'solve' it, back a
while ago at the time, was to go to the U2W cards, plus, make sure the Adaptec
BIOS on those U2W cards was at the latest revision level.

I've already given up on the 2940UW cards just as you noted above.

But there is more here, I think. Yes, the 29160 cards are working totally
stable as far as I can see on even the 320 level Seagate hard drives.

And .. for serious test purposes I've also just carried out some other research
here on a 915GAVL board with a U2W card in it. I've been able to complete
several complete SCSI DDS-3 4MM data cartridge backups on this box with no
errors with the U2W card. Now comes the interesting part.

I'd already replaced an earlier Adaptec card with the same 2940U2W version in a
Pentium II box with the same Seagate model SCSI DDS-3 4MM 23GB tape unit in it.
And I've even been having serious total system lockups red-light hard drive
type during total tape backups there. Curious, I looked at the SYSINFO 8.25
report on the driver profile on both it and the 915GAVL unit, which is giving
me no trouble with the U2W card there. AHA! There is a difference.

On the 915GAVL report, there are two additional functions showing in the
SYSINFO report for this card. They are "Memwrite+invalidate, System Errors",
which do *NOT* show up on the Pentium II system. As well, the 915GAVL data
shows a Cache Line Size of 64, which the Pentium II system shows a Cache Line
Size of 32.

Thinking about that, I chose to test the Pentium II box, for stability reasons,
with a command line instance of <cache /lazy:off> just after boot. I see
*MUCH* better stability in the Pentium II system with very, very heavy COM port
and IRQ use with that, together with very large data file writes on the HPFS
file partitions there with that change.

More important, testing the total backup to the SCSI DDS-3 4MM DAT drive on it
also finishes, as far as I can see, completely clean now with no lockups so
far. Interesting.

So .. with even MCP2 XRC05 and so on, it looks like maybe this is, in fact,
related to what you have posted, yet. The drive in this unit is also a level
320 Seagate unit.

As well, what is required now here, to perhaps really tell us what might be
important, courtesy of your post, is to slam a new 29160 card which I have into
this Pentium II box. Then do the tests for the whole stability and SCSI tape
backup without <cache /lazy:off>. That to see if we get out of the long SCSI
write snarls on it courtesy of the 29160 level card.

This takes hours of work at my level of use for this box. I'll get there in a
few days, hopefully. But I really need to know the right answer here, as may
be helpful to more than my doghouse around here, smile.
--
--> Sleep well; OS2's still awake! ;)

Mike Luther
Richard Steiner
2007-11-23 22:45:53 UTC
Permalink
On Tue, 20 Nov 2007 23:15:01 +0000 in comp.os.os2.bugs,
Post by Mike Luther
Also, there are I think three basic models of 2940 Adaptec cards.
The 2940UW, the 2940UWPro and per my recall the 2940U2W card.
There are a number of other models. The basic Adaptec AHA-2940 was a
50-pin PCI narrow SCSI card, and the 2940U was a narrow "Ultra-SCSI"
card that used 50-pin cables but also supported 40 MB/sec transfers.

Not relevant for most folks these days. This box is still a 50-pin
SCSI machine using a 2940U, though. It seems to work. :-)
--
-Rich Steiner >>>---> http://www.visi.com/~rsteiner >>>---> Mableton, GA USA
Mainframe/Unix bit twiddler by day, OS/2+Linux+DOS hobbyist by night.
WARNING: I've seen FIELDATA FORTRAN V and I know how to use it!
The Theorem Theorem: If If, Then Then.
Aidan Grey
2007-11-15 00:56:17 UTC
Permalink
Post by Mike Luther
I can't recall where I noticed this but want to follow up on it.
Somewhere I saw where it was posted that having an Adaptec SCSI device driver
card in an OS/2 box which did not connect with anything caused errors in the
OS/2 operations.
Can anyone re-visit or furnish data on this? Reason being is that I'd like to
create a generic OS/2 box for mobile drive tray debugging, backup and cloning
use that could handle both SATA drive trays and SCSI drive trays at the same
time. If I were to plug in a SATA drive tray with the appropriate drive but
not a SCSI tray, would this result in what was claimed?
As well, what if I were using a SCSI drive tray for the actual hard drive, but
no SATA drive tray in that slot or slots?
Thanks!
--
--> Sleep well; OS2's still awake! ;)
Mike Luther
I would suggest that in this case the Adaptec BASEDEV or DEVICE line should
follow any other mass storage lines in CONFIG.SYS. Otherwise, adding a drive
or other device might cause your drive letters to change.


Aidan Grey
William L. Hartzell
2007-11-17 03:02:14 UTC
Permalink
Post by Aidan Grey
Post by Mike Luther
I can't recall where I noticed this but want to follow up on it.
Somewhere I saw where it was posted that having an Adaptec SCSI device driver
card in an OS/2 box which did not connect with anything caused errors in the
OS/2 operations.
Can anyone re-visit or furnish data on this? Reason being is that I'd like to
create a generic OS/2 box for mobile drive tray debugging, backup and cloning
use that could handle both SATA drive trays and SCSI drive trays at the same
time. If I were to plug in a SATA drive tray with the appropriate drive but
not a SCSI tray, would this result in what was claimed?
As well, what if I were using a SCSI drive tray for the actual hard drive, but
no SATA drive tray in that slot or slots?
Thanks!
--
--> Sleep well; OS2's still awake! ;)
Mike Luther
I would suggest that in this case the Adaptec BASEDEV or DEVICE line should
follow any other mass storage lines in CONFIG.SYS. Otherwise, adding a drive
or other device might cause your drive letters to change.
Aidan Grey
If he is using LVM, then he could assign a drive letter to the removable
drive that it would have always, even when there is a conflict (it would
not show or be usable in that case). If he further used the kernel
command RESERVERDRIVELETTER=(some letter like S) in config.sys, then his
CD class drives would follow that letter. With proper forethought the
location of the Basedev, if after the basedev for the boot drive(s)
would not be a problem, he would not have moving drive letters.

In fact the only problem would be with the BIOS limit on Boot devices,
if by some chance he wanted to make the removeable bootable. Which is a
different can of worms. (And one which he did not ask.)

Mike as to your question, I don't do SCSI, and have no answer for you.
But from reviewing past posts, generally having no devices attached to
the SCSI chain causes the basedev for the adapter to not load. As SATA
is another name for IDE, that driver will load even with no devices
attached as long as it is enabled in BIOS (I've not tested this, so
grain of salt time).
--
Bill
Thanks a Million!
Will Honea
2007-11-14 06:48:22 UTC
Permalink
Post by Mike Luther
I can't recall where I noticed this but want to follow up on it.
Somewhere I saw where it was posted that having an Adaptec SCSI device
driver card in an OS/2 box which did not connect with anything caused
errors in the OS/2 operations.
Can anyone re-visit or furnish data on this? Reason being is that I'd
like to create a generic OS/2 box for mobile drive tray debugging, backup
and cloning use that could handle both SATA drive trays and SCSI drive
trays at the same
time. If I were to plug in a SATA drive tray with the appropriate drive
but not a SCSI tray, would this result in what was claimed?
As well, what if I were using a SCSI drive tray for the actual hard drive,
but no SATA drive tray in that slot or slots?
Mike, I've got PATA, SCSI, and SATA in this new Lenovo box here with only a
couple of piss-ant problems relating to drive enumeration by both the BIOS
and various OS's. I have no problems with missing drives on any
combination of the ports beyond the weird way the drives are enumerated.
The box came with only a SATA hard drive and a PATA DVD-RW. That made the
SATA drive drive 0. Plugged in a PATA disk and the old IDE enumeration
took over - the added drive was drive 0 and the SATA drive was bumped to
drive 1 as far as the BIOS was concerned. Added and actual SCSI drive to
the Adaptec adapter and the SATA drive got bumped to Drive 2 (PATA drive 0,
SCSI drive 1, SATA drive 2). The system is stable with any combo but this
musical chairs game with the drive position setup is enough to drive one to
drink - not to mention trying to figure out the boot sequence.
--
Will Honea
--
Posted via a free Usenet account from http://www.teranews.com
Loading...