17 Oct
2010
17 Oct
'10
6:44 a.m.
Right now I have two disks and one CD-ROM in my Alpha 433au (all attached to
the internal ribbon cable, if you have seen my question about the storage
shelf, that is not connected in this instance). The SRM recognises them
correctly thus:
dka0.0.0.1009.0 DKA0 COMPAQ ST32550W 6415
dka400.4.0.1009.0 DKA400 MATSHITA CD-ROM CR-508 XS03
dka600.6.0.1009.0 DKA600 IBM-PSG DDYS-T18350N M S96E
However, when I start the Debian installer, during the boot sequence it does
a SCSI reset and then only recognises the first disk, thus:
scsi(0:0): Resetting SCSI BUS
scsi0 : QLogic QLA1040 PCI to SCSI Host Adapter
Firmware version: 7.65.06, Driver version 3.26
Vendor: COMPAQ Model: ST32550W Rev: 6415
Type: Direct-Access ANSI SCSI revision: 02
So, for whatever reason, the CD-ROM and the other hard disk disappear. I
have checked the termination and it *seems* to be correct in that the disks
and the CD-ROM are not (or *should* not) terminating anything. There is a
terminator at the end of the SCSI cable, and I believe the QLA1040 SCSI
adapter does its own termination:
>>sh pk*
pka0_host_id 7
pka0_mode ultra
pka0_soft_term diff
Does anyone have any suggestions why the SCSI reset during boot should make
the CD-ROM and the other HDD disappear?
Thanks
Rob
17 Oct
17 Oct
10:17 a.m.
On 17 October 2010 11:44, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
The QLA1040 isn't differential is it? The soft_term should be just 'on' I
think.
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
>>sh pk*
pka0_host_id ? ? ? ? ? ?7
pka0_mode ? ? ? ? ? ? ? ultra
pka0_soft_term ? ? ? ? ?diff 
11:27 a.m.
At 3:17 PM +0100 10/17/10, Adrian Graham wrote:
On 17 October 2010 11:44, Rob Jarratt
<robert.jarratt at ntlworld.com> wrote:
The QLA1040 isn't differential is it? The soft_term should be just
'on' I think.
I think it's UW SCSI, it would help to know exactly what model card.
This might help.
http://www.dectrader.com/docs/set3/ucr_na-CYBEK_KZPBA_UG_B01.pdf-1.pdf
I didn't look at the manual, and I'm trying to remember what there
are for gotcha's. I think you can use the Narrow connector and *1*
UW SCSI connector.
Rob, if you haven't, try putting it in a different PCI slot.
Zane
--
| Zane H. Healy | UNIX Systems Administrator |
| healyzh at aracnet.com | OpenVMS Enthusiast |
| | Photographer |
+----------------------------------+----------------------------+
| My flickr Photostream |
| http://www.flickr.com/photos/33848088 at N03/ |
>>sh
pk*
pka0_host_id 7
pka0_mode ultra
pka0_soft_term diff
5:39 p.m.
The card is a KZPBA-CX. Web searches would indicate that this is really the
same as the -CA model. I tried it in two different PCI slots. In the
behaviour was exactly the same and in the other it failed to be recognised
by the SRM. Thanks for the link to the manual, I will study it carefully.
Regards
Rob
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Zane H. Healy
Sent: 17 October 2010 16:27
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: More SCSI Confusion
At 3:17 PM +0100 10/17/10, Adrian Graham wrote:
The QLA1040 isn't differential is it? The soft_term should be just 'on'
I think.
I think it's UW SCSI, it would help to know exactly what model card.
This might help.
http://www.dectrader.com/docs/set3/ucr_na-CYBEK_KZPBA_UG_B01.pdf-
1.pdf
I didn't look at the manual, and I'm trying to remember what there are for
gotcha's. I think you can use the Narrow connector and *1* UW SCSI
connector.
Rob, if you haven't, try putting it in a different PCI slot.
Zane
--
| Zane H. Healy | UNIX Systems Administrator |
| healyzh at aracnet.com | OpenVMS Enthusiast |
| | Photographer |
+----------------------------------+----------------------------+
| My flickr Photostream |
| http://www.flickr.com/photos/33848088 at N03/ |
On 17 October 2010 11:44, Rob Jarratt
<robert.jarratt at ntlworld.com>
wrote:
>>sh pk*
pka0_host_id 7
pka0_mode ultra
pka0_soft_term diff
18 Oct
18 Oct
7:21 a.m.
On 17 October 2010 22:39, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
The card is a KZPBA-CX. Web searches would indicate
that this is really the
same as the -CA model. I tried it in two different PCI slots. In the
behaviour was exactly the same and in the other it failed to be recognised
by the SRM. Thanks for the link to the manual, I will study it carefully.
I know the KZPBA-CX (and its differential cousin the CY) far too well
for my own good :) I've never successfully used both internal and
external ports even though the card itself is supposed to auto detect
and either do a soft term or not depending on what's connected to it.
I've had a LOT of problems with the internal SCSI cable on the 433 and
500, pka0_soft_term should just be 'on' so if you're having issues
it's either the card or the cable.
Other thing to check is that your devices aren't terminated
themselves, providing TERM PWR is fine but not actual termination.
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
12:48 p.m.
I suspected the cable a while back and bought a replacement, which had no
effect. I have just ordered another card too so I will see if changing it
will help when it arrives. None of the devices are terminated themselves as
far as I can tell, but when I replace the card I will check again. Also, I
now have pka0_soft_term set to "on" as you recommend.
Thanks
Rob
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Adrian Graham
Sent: 18 October 2010 12:22
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: More SCSI Confusion
On 17 October 2010 22:39, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
> The card is a KZPBA-CX. Web searches would indicate that this is
> really the same as the -CA model. I tried it in two different PCI
> slots. In the behaviour was exactly the same and in the other it
> failed to be recognised by the SRM. Thanks for the link to the manual, I
will
study it carefully.
I know the KZPBA-CX (and its differential cousin the CY) far too well for
my
own good :) I've never successfully used both
internal and external ports
even though the card itself is supposed to auto detect and either do a
soft
term or not depending on what's connected to it.
I've had a LOT of problems with the internal SCSI cable on the 433 and
500,
pka0_soft_term should just be 'on' so if
you're having issues it's either
the
card or the cable.
Other thing to check is that your devices aren't terminated themselves,
providing TERM PWR is fine but not actual termination.
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
17 Oct
17 Oct
11:32 a.m.
Yes I thought about that and I tried changing it to "on" but it did not make
a difference; and with the "diff" setting it does seem to work, at least
some of the time.
I also just found out something strange. If I remove any devices with a SCSI
ID numerically lower than the CD-ROM then both the CD-ROM and the higher
HDDs are still seen after the SCSI reset. To prove this I took out DKA600
and changed the SCSI ID of DKA0 to make it DKA500. This time, after the
reset, it recognised both the CD-ROM (ID 4) and the disk that was at ID 0
and now was at ID 5.
Regards
Rob
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Adrian Graham
Sent: 17 October 2010 15:17
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: More SCSI Confusion
On 17 October 2010 11:44, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
The QLA1040 isn't differential is it? The soft_term should be just 'on' I
think.
>>sh
pk*
pka0_host_id ? ? ? ? ? ?7
pka0_mode ? ? ? ? ? ? ? ultra
pka0_soft_term ? ? ? ? ?diff
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
5:33 p.m.
On 17/10/10 16:32, Rob Jarratt wrote:
Yes I thought about that and I tried changing it to
"on" but it did not make
a difference; and with the "diff" setting it does seem to work, at least
some of the time.
I also just found out something strange. If I remove any devices with a SCSI
ID numerically lower than the CD-ROM then both the CD-ROM and the higher
HDDs are still seen after the SCSI reset. To prove this I took out DKA600
and changed the SCSI ID of DKA0 to make it DKA500. This time, after the
reset, it recognised both the CD-ROM (ID 4) and the disk that was at ID 0
and now was at ID 5.
Regards
Rob
Might be a red herring Rob, but I seem to remember certainly for the
BA356 integrated into the Alphaserver 1000A that the enclosure could be
operated in a split bus mode where half the enclosure was on one bus and
the other half on another.
Just a thought (you might want to cross post this to comp.os.vms).
Mark.
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Adrian Graham
Sent: 17 October 2010 15:17
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: More SCSI Confusion
On 17 October 2010 11:44, Rob Jarratt<robert.jarratt at ntlworld.com> wrote:
The QLA1040 isn't differential is it? The soft_term should be just
'on' I
think.
>>
sh pk*
>>
pka0_host_id 7
pka0_mode ultra
pka0_soft_term diff
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
18 Oct
18 Oct
12:50 p.m.
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Mark Wickens
Sent: 17 October 2010 22:34
To: General Discussion: On-Topic Posts Only
Subject: Re: More SCSI Confusion
On 17/10/10 16:32, Rob Jarratt wrote:
think.
Might be a red herring Rob, but I seem to remember certainly for the
BA356 integrated into the Alphaserver 1000A that the enclosure could be
operated in a split bus mode where half the enclosure was on one bus and
the other half on another.
Just a thought (you might want to cross post this to comp.os.vms).
Mark.
I posted this to comp.sys.dec and someone told me there about the split bus.
I will check this.
Thanks
Rob
Yes I thought about that and I tried changing it
to "on" but it did
not make a difference; and with the "diff" setting it does seem to
work, at least some of the time.
I also just found out something strange. If I remove any devices with
a SCSI ID numerically lower than the CD-ROM then both the CD-ROM and
the higher HDDs are still seen after the SCSI reset. To prove this I
took out DKA600 and changed the SCSI ID of DKA0 to make it DKA500.
This time, after the reset, it recognised both the CD-ROM (ID 4) and
the disk that was at ID 0 and now was at ID 5.
Regards
Rob
> -----Original Message-----
> From: cctalk-bounces at classiccmp.org [mailto:cctalk-
> bounces at classiccmp.org] On Behalf Of Adrian Graham
> Sent: 17 October 2010 15:17
> To: General Discussion: On-Topic and Off-Topic Posts
> Subject: Re: More SCSI Confusion
>
> On 17 October 2010 11:44, Rob Jarratt<robert.jarratt at ntlworld.com>
wrote:
>>> sh pk*
>>>
pka0_host_id 7
pka0_mode ultra
pka0_soft_term diff
The QLA1040 isn't differential is it? The soft_term should be just
'on' I
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
7:13 p.m.
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Rob Jarratt
Sent: 18 October 2010 17:50
To: 'General Discussion: On-Topic and Off-Topic Posts'; 'General
Discussion:
On-Topic Posts Only'
Subject: RE: More SCSI Confusion
> -----Original Message-----
> From: cctalk-bounces at classiccmp.org [mailto:cctalk-
> bounces at classiccmp.org] On Behalf Of Mark Wickens
> Sent: 17 October 2010 22:34
> To: General Discussion: On-Topic Posts Only
> Subject: Re: More SCSI Confusion
>
> On 17/10/10 16:32, Rob Jarratt wrote:
> > Yes I thought about that and I tried changing it to "on" but it did
> > not make a difference; and with the "diff" setting it does seem to
> > work, at least some of the time.
> >
> > I also just found out something strange. If I remove any devices
> > with a SCSI ID numerically lower than the CD-ROM then both the
> > CD-ROM and the higher HDDs are still seen after the SCSI reset. To
> > prove this I took out DKA600 and changed the SCSI ID of DKA0 to make
it
DKA500.
Might be a red herring Rob, but I seem to remember certainly for the
BA356 integrated into the Alphaserver 1000A that the enclosure could
be operated in a split bus mode where half the enclosure was on one
bus and the other half on another.
Just a thought (you might want to cross post this to comp.os.vms).
Mark.
I posted this to comp.sys.dec and someone told me there about the split
bus.
I will check this.
Thanks
Rob
I have confirmed that it is indeed working in split bus mode. It seems I
need a BA35X-MF jumper module to get it to work in single bus mode.
Regards
Rob
This
time, after the reset, it recognised both the CD-ROM (ID 4) and
the disk that was at ID 0 and now was at ID 5.
Regards
Rob
> -----Original Message-----
> From: cctalk-bounces at classiccmp.org [mailto:cctalk-
> bounces at classiccmp.org] On Behalf Of Adrian Graham
> Sent: 17 October 2010 15:17
> To: General Discussion: On-Topic and Off-Topic Posts
> Subject: Re: More SCSI Confusion
>
> On 17 October 2010 11:44, Rob Jarratt<robert.jarratt at ntlworld.com>
wrote:
>>>> sh pk*
>>>>
> pka0_host_id 7
> pka0_mode ultra
> pka0_soft_term diff
>
The QLA1040 isn't differential is it? The soft_term should be just
'on' I
think.
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
19 Oct
19 Oct
5:23 a.m.
On 19 October 2010 00:13, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
That should be under the other fan, you normally just swap them over
top to bottom. If you look in the front of the enclosure you'll see a
pin poking through near the top and bottom, I think they're marked 'T'
and 'J'. This shows you which is currently the jumper and which is the
terminator.
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
Rob
I have confirmed that it is indeed working in split bus mode. It seems I
need a BA35X-MF jumper module to get it to work in single bus mode.
6:20 a.m.
As far as I can tell, I only have a terminator module (behind slot 6), but
not a jumper module.
Regards
Rob
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Adrian Graham
Sent: 19 October 2010 10:24
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: More SCSI Confusion
On 19 October 2010 00:13, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
That should be under the other fan, you normally just swap them over top
to
Rob
I have confirmed that it is indeed working in split bus mode. It seems
I need a BA35X-MF jumper module to get it to work in single bus mode.
bottom. If you look in the front of the enclosure
you'll see a pin poking
through near the top and bottom, I think they're marked 'T'
and 'J'. This shows you which is currently the jumper and which is the
terminator.
--
adrian/witchy
Owner of Binary Dinosaurs, the UK's biggest home computer collection?
www.binarydinosaurs.co.uk
5002
days inactive
5004
days old
11 comments
4 participants
participants (4)
-
binarydinosaurs＠gmail.com -
healyzh＠aracnet.com -
mark＠wickensonline.co.uk -
robert.jarratt＠ntlworld.com