Discussion:
P400 on DS10/DS15
(too old to reply)
John H. Reinhardt
2020-04-10 00:07:32 UTC
Permalink
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
\\
I'm getting close to trialing a SATA SSD solution for RX2600 units w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter plus cabling). The SATA drives will be deployed in an external enclosure. I have a bill of materials drafted. Parts are trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires 3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss. From the Technical Summary documents on both, it's a PCI V2.1 compliant bus supporting either %.0V or 3.3V PCI cards.

But all the PICe to PCI converter boards I see say they are PCI V2.3 compliant so I don't know if they would work with the DS10/15 PCI bus.

This one is common on Amazon in the US StarTech PCIPEX1 - <https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>

Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a P400 or P800 series SATA HBA?
--
John H. Reinhardt
Rod Regier
2020-04-10 10:09:08 UTC
Permalink
Eberhard Heuser or Stephen Hoffman are the subject matter experts on device boot support that I follow. Hopefully they'll step into this thread.

I'm still waiting on my PCIx to PCIe adapter to be delivered so I'm not certain about PCIX slot hardware comparability yet.

I can tell you that I tried the PCIX compatible P600 card on a DS15.
It could use the attached storage for data, but there was no apparent DS15 firmware support for booting from it. Because of that issue I doubt a P400/p800/P410 can be booted on a DS15 (or a DS10) even if the buss compatibility issues can be resolved with an adapter :-(

The P410i has much better pricing/availability and is from the same driver support family as the P400 or P800.
John H. Reinhardt
2020-04-11 03:05:22 UTC
Permalink
Post by Rod Regier
Eberhard Heuser or Stephen Hoffman are the subject matter experts on device boot support that I follow. Hopefully they'll step into this thread.
I'm still waiting on my PCIx to PCIe adapter to be delivered so I'm not certain about PCIX slot hardware comparability yet.
I can tell you that I tried the PCIX compatible P600 card on a DS15.
It could use the attached storage for data, but there was no apparent DS15 firmware support for booting from it. Because of that issue I doubt a P400/p800/P410 can be booted on a DS15 (or a DS10) even if the buss compatibility issues can be resolved with an adapter :-(
The P410i has much better pricing/availability and is from the same driver support family as the P400 or P800.
I can get a P600 card with 256MB cache for under $30 on Ebay. Is there any benefit to the P410 w/ PCI-PCIe adapter vs the P600 in a DS10? (assuming it works, I think it will) for data? I know I can't boot from it.
--
John H. Reinhardt
Eberhard Heuser
2020-04-11 07:24:42 UTC
Permalink
Post by Rod Regier
Eberhard Heuser or Stephen Hoffman are the subject matter experts on
device boot support that I follow. Hopefully they'll step into this
thread.
I'm still waiting on my PCIx to PCIe adapter to be delivered so I'm
not certain about PCIX slot hardware comparability yet.
I can tell you that I tried the PCIX compatible P600 card on a DS15.
It could use the attached storage for data, but there was no apparent
DS15 firmware support for booting from it.  Because of that issue I
doubt a P400/p800/P410 can be booted on a DS15 (or a DS10) even if
the buss compatibility issues can be resolved with an adapter :-(
The P410i has much better pricing/availability and is from the same
driver support family as the P400 or P800.
I can get a P600 card with 256MB cache for under $30 on Ebay.  Is
there any benefit to the P410 w/ PCI-PCIe adapter vs the P600 in a
DS10? (assuming it works, I think it will) for data?  I know I can't
boot from it.
Don't be surprised that the performance of such a solution is not that good.

A P410 is a real step forward because the device uses 6 GB/s instead of
3 GB/s and the usage of a (sata-)SSD.

Eberhard
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
John H. Reinhardt
2020-04-11 08:10:32 UTC
Permalink
Post by Eberhard Heuser
Eberhard Heuser or Stephen Hoffman are the subject matter experts on device boot support that I follow. Hopefully they'll step into this thread.
I'm still waiting on my PCIx to PCIe adapter to be delivered so I'm not certain about PCIX slot hardware comparability yet.
I can tell you that I tried the PCIX compatible P600 card on a DS15.
It could use the attached storage for data, but there was no apparent DS15 firmware support for booting from it.  Because of that issue I doubt a P400/p800/P410 can be booted on a DS15 (or a DS10) even if the buss compatibility issues can be resolved with an adapter :-(
The P410i has much better pricing/availability and is from the same driver support family as the P400 or P800.
I can get a P600 card with 256MB cache for under $30 on Ebay.  Is there any benefit to the P410 w/ PCI-PCIe adapter vs the P600 in a DS10? (assuming it works, I think it will) for data?  I know I can't boot from it.
Don't be surprised that the performance of such a solution is not that good.
A P410 is a real step forward because the device uses 6 GB/s instead of 3 GB/s and the usage of a (sata-)SSD.
Eberhard
I don't expect the same performance as if it were in a more modern system. The P600 has a PCI-X bus at up to 132Mhz but the DS10's best is 66MHz. The P600 has a theoretical speed of about 1 GB/s which it won't hit on the DS10's PCI bus. The P410 has a higher speed which also won't be used. But I would expect (hope) it to compare favorably to the best SCSI choice which is something on the order of an Ultra 160 SCSI HBA such as the 3X-KZPEA-DB (Adaptec 39160) having a bandwidth of about 160MB/s.

Does that seem reasonable?
--
John H. Reinhardt
Rod Regier
2020-04-10 10:15:22 UTC
Permalink
This is the adapter I have on order. It has been tested with OpenVMS by Eberhard Heuer.
https://www.ebay.com/itm/Converter-Adapter-Pcie-16x-for-PCI-32-Bit-Chipset-Plx-PEX8112/293081364769
Eberhard Heuser
2020-04-10 10:36:19 UTC
Permalink
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units
w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe
adapter plus cabling).  The SATA drives will be deployed in an
external enclosure.  I have a bill of materials drafted. Parts are
trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires
3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the Technical
Summary documents on both, it's a PCI V2.1 compliant bus supporting
either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3
compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 -
<https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a
P400 or P800 series SATA HBA?
I doubt that the SRM console will offer the P400-controller as a boot
device.

Eberhard
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
John H. Reinhardt
2020-04-10 14:27:53 UTC
Permalink
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter plus cabling).  The SATA drives will be deployed in an external enclosure.  I have a bill of materials drafted. Parts are trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires 3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the Technical Summary documents on both, it's a PCI V2.1 compliant bus supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3 compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 - <https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a P400 or P800 series SATA HBA?
I doubt that the SRM console will offer the P400-controller as a boot device.
Eberhard
That's pretty much what I figured. I might get a StarTech card since hey are USA based and see if it works for data.
--
John H. Reinhardt
John H. Reinhardt
2020-04-12 17:59:34 UTC
Permalink
This post might be inappropriate. Click to display it.
David Turner
2020-04-14 20:07:32 UTC
Permalink
I have BRAND NEW 300GB 15K U320 3.5" SCSI drives in stock
These are 80 pin but I do have 80 to 68pin converters which work
perfectly with no loss

Link:
https://www.islandco.com/3R-A6726-AA-hp-300gb-15k-u320-universal-hot-plug-hard-drive-scsi



David
island computers
Post by Eberhard Heuser
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units
w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe
adapter plus cabling).  The SATA drives will be deployed in an
external enclosure.  I have a bill of materials drafted. Parts are
trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires
3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the Technical
Summary documents on both, it's a PCI V2.1 compliant bus supporting
either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3
compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 -
<https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a
P400 or P800 series SATA HBA?
I doubt that the SRM console will offer the P400-controller as a boot
device.
Eberhard
Rod Regier
2020-04-13 18:33:52 UTC
Permalink
P600 uses two cache batteries, same as the SA6402.
One of the SKUs is 307132-001, 3rd party SKU is CS-RAC6400SL

Vendors actually selling true "new" batteries are elusive.

The NiMH batteries have a rated life of no more than 4 years, and an HP recommendation for proactive replacement on a 3-year cycle.

HP mfgr batteries have a YY/MM date code marked on the edge of the batteries.
Scott Dorsey
2020-04-13 21:16:31 UTC
Permalink
Post by Rod Regier
P600 uses two cache batteries, same as the SA6402.
One of the SKUs is 307132-001, 3rd party SKU is CS-RAC6400SL
Vendors actually selling true "new" batteries are elusive.
The NiMH batteries have a rated life of no more than 4 years, and an HP recommendation for proactive replacement on a 3-year cycle.
HP mfgr batteries have a YY/MM date code marked on the edge of the batteries.
It is not too hard to add a battery holder that will permit use of
interchangeable NiMH cells. Much better than those weird flat lithium
batteries used in some equipment.

If you don't want to do this, take your existing battery to be recelled.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Rod Regier
2020-04-13 18:37:59 UTC
Permalink
The biggest thruput bottleneck of any of these variant solutions is the speed of the PCIX slot :-(

Which is why the industry created multiple speed tiers of PCIe and new mainboards were designed to support PCIe...

But not everyone has the resources to update to the latest and greatest :-(
John H. Reinhardt
2020-04-13 18:54:10 UTC
Permalink
Post by Rod Regier
The biggest thruput bottleneck of any of these variant solutions is the speed of the PCIX slot :-(
Which is why the industry created multiple speed tiers of PCIe and new mainboards were designed to support PCIe...
But not everyone has the resources to update to the latest and greatest :-(
True. I mean, it's an EV67 Alpha at 617Mhz so you can't expect it to run like a 2.5Ghz Itanium or a 3.93GHz Xeon. I just want something that will perform in the same general area that the typical PCI SCSI device for that era would be expected, just with newer drives that aren't close to EOL.
--
John H. Reinhardt
Rod Regier
2020-04-13 18:50:27 UTC
Permalink
I've used the P600 on a DS15 with SSD SATA drive(s) as a data disk controller.
Fresh batteries so caching was enabled. Burned in using UETP, no issues.

Again I was following the wisdom of Eberhard Heuser in deploying the pieces.

The magic post-boot driver spell:

Add the following lines to file sys$system:SYS$USER_CONFIG.DAT
device = "Smart Array P600"
name = PK
driver = sys$pkrdriver.exe
adapter = PCI
id = 0x3220103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT, BOOT
end_device

Next reboot configured logical volumes (RAID, JBOD) will be visible.

n.b. MSA$UTIL is used to manage the P600 online, or ORCA offline

MSA$UTIL can also be used to determine current battery charge state.
If the batteries haven;t reached a charge in 3 power-up hours, they have an issue.
Likely culprits in order of most to least likely:
EOL batteries
Battery contact to card issue
Card battery circuit issue (rare).

I like using a nylon scratch brush to refurb used card contacts before attaching batteries and plugging into slots.

Unless you abuse it a single nylon scratch brush will last you a lifetime.

Just don't put it in a garment pocket - that's like placing glass wool there :-(

https://www.amazon.com/Jewelers-Nylon-Fiberglass-Scratch-Brush/dp/B079YVJHG3/ref=sr_1_21?dchild=1&keywords=nylon+scratch+brush&qid=1586803642&sr=8-21

If you are using a serial console "F1" is generated by typing "ESC" "1" etc.
John H. Reinhardt
2020-04-13 19:00:03 UTC
Permalink
Post by Rod Regier
I've used the P600 on a DS15 with SSD SATA drive(s) as a data disk controller.
Fresh batteries so caching was enabled. Burned in using UETP, no issues.
Again I was following the wisdom of Eberhard Heuser in deploying the pieces.
Add the following lines to file sys$system:SYS$USER_CONFIG.DAT
device = "Smart Array P600"
name = PK
driver = sys$pkrdriver.exe
adapter = PCI
id = 0x3220103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT, BOOT
end_device
Next reboot configured logical volumes (RAID, JBOD) will be visible.
n.b. MSA$UTIL is used to manage the P600 online, or ORCA offline
MSA$UTIL can also be used to determine current battery charge state.
If the batteries haven;t reached a charge in 3 power-up hours, they have an issue.
EOL batteries
Battery contact to card issue
Card battery circuit issue (rare).
I like using a nylon scratch brush to refurb used card contacts before attaching batteries and plugging into slots.
Unless you abuse it a single nylon scratch brush will last you a lifetime.
Just don't put it in a garment pocket - that's like placing glass wool there :-(
https://www.amazon.com/Jewelers-Nylon-Fiberglass-Scratch-Brush/dp/B079YVJHG3/ref=sr_1_21?dchild=1&keywords=nylon+scratch+brush&qid=1586803642&sr=8-21
If you are using a serial console "F1" is generated by typing "ESC" "1" etc.
It's good to know that MSA$UTIL will still manage these cards. Makes it easier. I will be using the serial console and I knew there were equivalents for "F1", etc. Thanks for the reminder.

The card it supposedly NOS and still sealed in the anti=static bag, but I don't hold much hope for the battery(ies) (the picture showed only one but it was hard to tell through the bag). It shipped today (Monday 4/13) from NYC, headed to Fort Worth via FedEx Ground so it will be a week or so before I get it. The other parts are arriving via Amazon - some are already here.
--
John H. Reinhardt
Rod Regier
2020-04-13 19:08:17 UTC
Permalink
correction - my P600 testing was on an RX2600. Having said that,
as long as you can get the p600 to plug into the PCIx slot is suspect you can get it to work as a data drive controller on a DS10 or DS15.

I've run the SA6402 on DS15s successfully, the P600 is the same design vintage.

Steven Hoffman much earlier authored a note on his web site on how to get a SA6402 working for boot on a DS15.

You might find this thread interesting:

https://groups.google.com/forum/#!topic/comp.os.vms/PAO0gcrKdOY
Rod Regier
2020-04-13 19:14:59 UTC
Permalink
My general experience with purchasing using HP option cards is generally good.
Expired batteries and legacy firmware are the primary issues.

I have the materials and cookbooks to upgrade firmware for:
SA6402
P400
P410
LSISAS1068

Don't know about recommended P600 firmware version or upgrade materials.

Mr. Google can probably help you on version details.

ORCA displays the firmware version, and it can also be displayed from MSA$UTIL.
Rod Regier
2020-04-13 19:18:21 UTC
Permalink
If you get a bench test to work to your satisfaction, you're going to have to sort out cabling/external enclosure for the SFF SATA devices.

There is no "nice" space in a DS10 to mount and power SFF SATA drives.

I've done most of that research for the RX2600 on how to externally cable and enclosure support SFF SATA drives. If you get to that point I would be happy to share that info.
John H. Reinhardt
2020-04-13 20:55:23 UTC
Permalink
Post by Rod Regier
correction - my P600 testing was on an RX2600. Having said that,
as long as you can get the p600 to plug into the PCIx slot is suspect you can get it to work as a data drive controller on a DS10 or DS15.
I've run the SA6402 on DS15s successfully, the P600 is the same design vintage.
Oh, that's quite a difference. lol That said, I have a SA6402 that "worked" on my DS10. I use the quotes because, the DS10 saw the card and even let mu upgrade the firmware using the DS10 V7.3-1 firmware disk, but I could not get it to see any drives on the SCSI buss. Even though when the cable was unplugged from the SA6402 and plugged into tow other SCSI cards, they saw the drives (a KZPBA-CY QLogic and a 3X-KZPEA-DB Adaptec 39160). I had problems getting the BIOS/ORCA to run with the serial console possibly because I forgot the ESC equivalents for the "Fx" keys. Hopefully this goes better. If there is a problem with the serial console running the BIOS then I may try putting a video card in and using the graphical console.

I had the same problem with a SA5302A which also should have worked so it was probably something I did wrong.
Post by Rod Regier
If you get a bench test to work to your satisfaction, you're going to have to sort out cabling/external enclosure for the SFF SATA devices.
There is no "nice" space in a DS10 to mount and power SFF SATA drives.
I've done most of that research for the RX2600 on how to externally cable and enclosure support SFF SATA drives. If you get to that point I would be happy to share that info.
Hopefully no external enclosures. I have this from Amazon: <https://www.amazon.com/gp/product/B01M0BIPYC> which should go into one of the DS10's 5.25" bays, probably right below the CD drive. And these cables to run from the P600 to the drive cage. <https://www.amazon.com/gp/product/B01CSUWHX2> Then one SCSI disk below the floppy (which I've exchanged for a SCSI2SD V6 disk emulator <http://www.codesrc.com/mediawiki/index.php?title=SCSI2SD> ) and I should be set.
--
John H. Reinhardt
David Turner
2020-04-14 20:11:38 UTC
Permalink
To get the SA6402/6404 working on a DS15 as a bootable device
set heap_expand 2MB (4mb for 6404)
set bootbios pya0
Init
correction - my P600 testing was on an RX2600.  Having said that,
as long as you can get the p600 to plug into the PCIx slot is suspect
you can get it to work as a data drive controller on a DS10 or DS15.
I've run the SA6402 on DS15s successfully, the P600 is the same design vintage.
Oh, that's quite a difference.  lol  That said, I have a SA6402 that
"worked" on my DS10.  I use the quotes because, the DS10 saw the card
and even let mu upgrade the firmware using the DS10 V7.3-1 firmware
disk, but I could not get it to see any drives on the SCSI buss.  Even
though when the cable was unplugged from the SA6402 and plugged into tow
other SCSI cards, they saw the drives (a KZPBA-CY QLogic and a
3X-KZPEA-DB Adaptec 39160).  I had problems getting the BIOS/ORCA to run
with the serial console possibly because I forgot the ESC equivalents
for the "Fx" keys.  Hopefully this goes better.  If there is a problem
with the serial console running the BIOS then I may try putting a video
card in and using the graphical console.
I had the same problem with a SA5302A which also should have worked so
it was probably something I did wrong.
If you get a bench test to work to your satisfaction, you're going to
have to sort out cabling/external enclosure for the SFF SATA devices.
There is no "nice" space in a DS10 to mount and power SFF SATA drives.
I've done most of that research for the RX2600  on how to externally
cable and enclosure support SFF SATA drives.  If you get to that point
I would be happy to share that info.
<https://www.amazon.com/gp/product/B01M0BIPYC> which should go into one
of the DS10's 5.25" bays, probably right below the CD drive.  And these
cables to run from the P600 to the drive cage.
<https://www.amazon.com/gp/product/B01CSUWHX2>  Then one SCSI disk below
the floppy (which I've exchanged for a SCSI2SD V6 disk emulator
<http://www.codesrc.com/mediawiki/index.php?title=SCSI2SD> ) and I
should be set.
Rod Regier
2020-04-14 11:48:49 UTC
Permalink
I'm onsite at the office today, so I can look at some hardware.

The P600 (and SA6402) PCIx card keying is *not* compatible with the DS10 PCIx slots.
The P600 is compatible with the RX2600 PCIx slots.
Since the SA6402 is keyed the same as the P600 they are both compatible with DS15 PCIx slots. I believe the DS15 PCIx is keyed differently from the DS10 because the DS15 also supports 3.3V PCIx options, which the DS10 does not :-(


The PLX8112-based PCIe to PCIx adapter I purchased *is* keying compatible with DS10, RX2600 PCIx slots.

US Ebay source for adapter I purchased from France:

https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013
John H. Reinhardt
2020-04-14 15:53:19 UTC
Permalink
Post by Rod Regier
I'm onsite at the office today, so I can look at some hardware.
The P600 (and SA6402) PCIx card keying is *not* compatible with the DS10 PCIx slots.
The P600 is compatible with the RX2600 PCIx slots.
Since the SA6402 is keyed the same as the P600 they are both compatible with DS15 PCIx slots. I believe the DS15 PCIx is keyed differently from the DS10 because the DS15 also supports 3.3V PCIx options, which the DS10 does not :-(
The PLX8112-based PCIe to PCIx adapter I purchased *is* keying compatible with DS10, RX2600 PCIx slots.
https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013
Looks like you are correct. The DS10 Technical documentation says that the PCI slots support both 3.3V and 5V options, the problem is that the SA6400 is 3.3V ONLY. The PCI connector does not have the proper keying to let it fit into a PCI slot that supplies both 3.3V and 5V. The DS10 PCI slots are keyed for both. So I mis-remembered trying the SA6402. I checked with the SA5302A and it has the proper keying on the PCI connector to allow it to fit into the DS10 where the SA6402 does not.

Looking at the P600 Quickspecs I see it is keyed the same as the SA6400 so it will not work also. Ah, well.

I will have to try the PCI-to-PCIe adapter and a P410
--
John H. Reinhardt
David Turner
2020-04-14 20:14:24 UTC
Permalink
Is there not an equivalent Proliant that may use a PCI-x/PCi-e riser card?
What I mean is, is that thje Proliants and Integrity servers share a lot
of hardware. I wonder if say a Proliant G4 might use a riser card that
DOES have PCI-e that could be stuffed into an rx2600
Post by Rod Regier
I'm onsite at the office today, so I can look at some hardware.
The P600 (and SA6402) PCIx card keying is *not* compatible with the DS10 PCIx slots.
The P600 is compatible with the RX2600 PCIx slots.
Since the SA6402 is keyed the same as the P600 they are both
compatible with DS15 PCIx slots.  I believe the DS15 PCIx is keyed
differently from the DS10 because the DS15 also supports 3.3V PCIx
options, which the DS10 does not :-(
The PLX8112-based PCIe to PCIx adapter I purchased *is* keying
compatible with DS10, RX2600 PCIx slots.
https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013
Looks like you are correct.  The DS10 Technical documentation says that
the PCI slots support both 3.3V and 5V options, the problem is that the
SA6400 is 3.3V ONLY.  The PCI connector does not have the proper keying
to let it fit into a PCI slot that supplies both 3.3V and 5V.  The DS10
PCI slots are keyed for both.  So I mis-remembered trying the SA6402.  I
checked with the SA5302A and it has the proper keying on the PCI
connector to allow it to fit into the DS10 where the SA6402 does not.
Looking at the P600 Quickspecs I see it is keyed the same as the SA6400
so it will not work also.  Ah, well.
I will have to try the PCI-to-PCIe adapter and a P410
John H. Reinhardt
2020-04-18 02:55:43 UTC
Permalink
Post by Rod Regier
I'm onsite at the office today, so I can look at some hardware.
The P600 (and SA6402) PCIx card keying is *not* compatible with the DS10 PCIx slots.
The P600 is compatible with the RX2600 PCIx slots.
Since the SA6402 is keyed the same as the P600 they are both compatible with DS15 PCIx slots.  I believe the DS15 PCIx is keyed differently from the DS10 because the DS15 also supports 3.3V PCIx options, which the DS10 does not :-(
The PLX8112-based PCIe to PCIx adapter I purchased *is* keying compatible with DS10, RX2600 PCIx slots.
https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013
Looks like you are correct.  The DS10 Technical documentation says that the PCI slots support both 3.3V and 5V options, the problem is that the SA6400 is 3.3V ONLY.  The PCI connector does not have the proper keying to let it fit into a PCI slot that supplies both 3.3V and 5V.  The DS10 PCI slots are keyed for both.  So I mis-remembered trying the SA6402.  I checked with the SA5302A and it has the proper keying on the PCI connector to allow it to fit into the DS10 where the SA6402 does not.
Looking at the P600 Quickspecs I see it is keyed the same as the SA6400 so it will not work also.  Ah, well.
I will have to try the PCI-to-PCIe adapter and a P410
All the parts are in house now except the PCI-to-PCIe adapter card. Ordered Tuesday on Ebay the seller (supposedly in Exeter, NH but apparently really in the UK or maybe China - their feedback is telling) but not even shipped yet according to Ebay. I found the card on Amazon but all sellers are outside the US (probably in China) and all have an estimated delivery date of May 9 to June 1st. The StarTech card which Amazon could have to me by Sunday only as a 1x PCI-e slot on the top and will not accommodate the 8x PCI-e tab of the P410.

So I guess I wait.

Meanwhile, I work with SCSI2SD cards for my two DS10's and MicroVAX 3100. And replace RIFA line filtering caps on my 3 BA23 power supplies and get my PDP-11/53 running (It gets a SCSI2SD card as well). And I could write bash scripts to standardize my Raspberry Pi 3B+ and 4B Group which will be running SimH and KLH10. Lots to do.
--
John H. Reinhardt
Rod Regier
2020-04-14 12:12:52 UTC
Permalink
I've successfully used offline ORCA w/DS15 and RX2600 using serial console with the alternate keymapping.

The OpenVMS O/S can't "see" any of the attached drives on the SA6402 (and prob P600) until they have been initialized with ORCA or MSA$UTIL.

The drives units (comprised of member disks) are configured as RAID sets or JBODs.
Rod Regier
2020-04-14 16:02:18 UTC
Permalink
How update P410 firmware to current (6.64) on an Integrity server:

https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-a17b1b7dcc284cf5a21ecfa7c6

Sorry, no comparable article for an Alpha server.
Eberhard Heuser
2020-06-14 10:10:19 UTC
Permalink
Post by Rod Regier
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-a17b1b7dcc284cf5a21ecfa7c6
Sorry, no comparable article for an Alpha server.
Some news in respect to the P410-controller:

1. You'll need a SAAP-License if you want to use Raid6. HP sells these licenses,
but they are very, very expensive. Otherwise you only can use Raid5.
2. I have successfully connected P2000-Box with a AJ751A G3 Controller with a
P410-Controller. SFF8087<->SFF8088 + SFF-8088<_>SFF8470 cables are needed,
too.

The advantage of this solution: 12 disk slots or even more(?) !

I'm considering to use a AP844B 6G SAS that you'll get cheap.
Rod Regier
2020-04-17 20:34:20 UTC
Permalink
Post by David Turner
Is there not an equivalent Proliant that may use a PCI-x/PCi-e riser card?
What I mean is, is that thje Proliants and Integrity servers share a lot
of hardware. I wonder if say a Proliant G4 might use a riser card that
DOES have PCI-e that could be stuffed into an rx2600
I don't have much knowledge of older Proliant internals.
I'm not sure if it is even technically feasible to some how graft on a PCIe buss riser card onto a mainboard that wasn't designed to support PCIe.

The other issue is that non-industry standard card buss components typically
aren't designed to mechanically interchange between product families
even from the same manufacturer. (i.e. probably some compatible PCIx cards but not internals).

I doubt there would be a mechanically compatible Proliant internal option that could substitute for the PCIx only riser in the RX2600 :-(

The RX2600 was released around 2003, so you would be searching for 2003 vintage Proliant components even if they were compatible :-(
Rod Regier
2020-04-18 16:54:53 UTC
Permalink
This PLX8112 adapter SKU that Eberhard and I have successfully used has 297 pieces in Exeter, NH stock:

https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013

I bought my first PLX8112 from Ebay French vendor suggested by Eberhard (E is in Germany).

My subsequent PLX8112 purchases will be from the above US Ebay contact if I am in a hurry, AliExpress if not.
John H. Reinhardt
2020-04-18 17:35:18 UTC
Permalink
Post by Rod Regier
https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013
I bought my first PLX8112 from Ebay French vendor suggested by Eberhard (E is in Germany).
My subsequent PLX8112 purchases will be from the above US Ebay contact if I am in a hurry, AliExpress if not.
Yeah. That's the same vendor. Looking at the Ebay listing it says Exeter, NH aas the location, but when you look at the seller info is says "Based in United Kingdom, uksf18 has been an eBay member since Jun 28, 2018" and if you read some of the negative feedback (23 over the last month vs 239 positive) you see numerous entries about not ever shipping the item and one or two remarking that the item shipped from LA or stuck in customs in Canada. They probably don't keep stock but drop ship from various places including AliExpress. I asked them on the 17th if they knew when they would ship and got a reply this morning "Sorry,as some reason we have delay the shipment, we will send it out on Monday, you can get it next week, ok?" so if they really do ship it Monday and it really does come from NH then I should get it by Thursday or Friday next week.

I saw the French vendor too (either Ebay or Amazon, I forget) but their estimate for delivery was in June. I'm just spoiled by the "normal" fast delivery times of Amazon.
--
John H. Reinhardt
John H. Reinhardt
2020-04-21 06:52:48 UTC
Permalink
Post by Rod Regier
https://www.ebay.ca/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013
I bought my first PLX8112 from Ebay French vendor suggested by Eberhard (E is in Germany).
My subsequent PLX8112 purchases will be from the above US Ebay contact if I am in a hurry, AliExpress if not.
Yeah. That's the same vendor.  Looking at the Ebay listing it says Exeter, NH aas the location, but when you look at the seller info is says "Based in United Kingdom, uksf18 has been an eBay member since Jun 28, 2018" and if you read some of the negative feedback (23 over the last month vs 239 positive) you see numerous entries about not ever shipping the item and one or two remarking that the item shipped from LA or stuck in customs in Canada.  They probably don't keep stock but drop ship from various places including AliExpress.  I asked them on the 17th if they knew when they would ship and got a reply this morning "Sorry,as some reason we have delay the shipment, we will send it out on Monday, you can get it next week, ok?" so if they really do ship it Monday and it really does come from NH then I should get it by Thursday or Friday next week.
I saw the French vendor too (either Ebay or Amazon, I forget) but their estimate for delivery was in June.  I'm just spoiled by the "normal" fast delivery times of Amazon.
The vendor (uksf18) replied today (1:30AM Monday CDT - further showing they aren't really in the US) that "Sorry,as some reason we have delay the shipment, we will send it out this week, you can get it in 5-7 working days, ok?" I replied to them that if they don't ship it such that I receive it by April 29th (the last day Ebay said it should arrive) then I would file with Ebay and Paypal for a refund and log a negative feedback.

I ordered two cards from the French vendor. Estimated delivery is April 29th to May 8th. I bet it still beats the one from Exeter.
--
John H. Reinhardt
Rod Regier
2020-04-18 16:59:24 UTC
Permalink
Sorry, didn't catch the Exeter reference in your Ebay vendor post.

The French Ebay vendor stocks in France, but had to do a dacge with his normal warehouse (Amazon) to get the card released ho he could ship it himself.
That stretched my delivery from him.
Rod Regier
2020-04-18 17:09:31 UTC
Permalink
FWIW, I have completed my RX2660 testing of the P410i.
These RX2660 combos with SATA SSD all worked:
P410i in plain PCIe slot
P410i in P400 PCIe slot
P410i plus PLX8112 adapter in PCIX slot

I tested with SATA breakout cable and different multiport cable to the front-access hot swap drive bays.

I had to update the boot pointers for each case, which wasn't startling.

I loaded the SATA SSD drive with patched V/I64 8.4 once and used it on all the variants w/o change. (Since ultimately the same P410i controller was talking to the same drive).

\\

Next up is testing on RX2600 with P410i plus PLX8112 adapter.
Since it is a bench-only test I'll use the SATA breakout cable to an external SATA drive and an external power source for the SATA SSD drive.

\\

When using the P410i with the PLX8112 I had to:
remove the PCI and heat shield parts from the P410i.
Run a edge wrap of insulating tape on both circuit card sides on the PCI bracket end of the P410i to avoid the PLX8112 PCI bracket from possibly shorting the
components and traces on the extreme end of the P410i card.
John H. Reinhardt
2020-04-18 17:43:22 UTC
Permalink
Post by Rod Regier
FWIW, I have completed my RX2660 testing of the P410i.
P410i in plain PCIe slot
P410i in P400 PCIe slot
P410i plus PLX8112 adapter in PCIX slot
I tested with SATA breakout cable and different multiport cable to the front-access hot swap drive bays.
I had to update the boot pointers for each case, which wasn't startling.
I loaded the SATA SSD drive with patched V/I64 8.4 once and used it on all the variants w/o change. (Since ultimately the same P410i controller was talking to the same drive).
\\
Next up is testing on RX2600 with P410i plus PLX8112 adapter.
Since it is a bench-only test I'll use the SATA breakout cable to an external SATA drive and an external power source for the SATA SSD drive.
\\
remove the PCI and heat shield parts from the P410i.
Run a edge wrap of insulating tape on both circuit card sides on the PCI bracket end of the P410i to avoid the PLX8112 PCI bracket from possibly shorting the
components and traces on the extreme end of the P410i card.
I have a RX2660 but something has gone wrong on the mainboard and I can't talk to it past the management console. It won't boot and won't do anything but let me run the console commands. I have to work more with that to see what happened. I had it powered off for about 18 months. Before it worked great, after, not so much. I'm hoping maybe it's the battery or something simple. That's on the list for the near future.

I also have a ZX6000 workstation version of the RZ2600. I was thinking of checking to see what it would take. I'm not sure if the PCI bus in it is PCI or PCI-e. I think it's PCI-X based on what you say. I'd have to go with an external housing for the SATA drives though as the 3-bay internal is not suitable for putting a SATA drive cage in it like the DS10 is.

BTW, that ICY-DOCK 2x3 drive cage fits perfectly into either 5.25" bay on the DS10. I put it on the top and moved the CD drive to the bottom. I'm replacing the CD drive with a SCSI DVD-ROM reader since the shipped CD is IDE and the IDE on the DS10 is dog slow - slower even than the CD drive.
--
John H. Reinhardt
Rod Regier
2020-04-18 18:07:30 UTC
Permalink
FWIW, Eberhard was using a ZX2000 in his OpenVMS/I64 experiments withe the P410i and the PLX8112.

Suggest digging up the Quickspecs for your specific ZX series model.

I suspect if it has an I64 processor and either a PCIx or PCIe buss
you could be in business.

Guess I'll cross off the Exeter vendor as a future option.

I've done some HW maint on the RX2660 but I didn't start out with
once as broken as yours.

I would be using the embedded ILO (System Event Log is your friend) and front panel lights to try and get a sense of what might be the lowest level issue.
hanging it. The iLO is a co-processor - it will work as long as the chassis has power.
John H. Reinhardt
2020-04-18 18:52:36 UTC
Permalink
Post by Rod Regier
FWIW, Eberhard was using a ZX2000 in his OpenVMS/I64 experiments withe the P410i and the PLX8112.
Suggest digging up the Quickspecs for your specific ZX series model.
Got them. Just haven't checked yet. I'm pretty sure it's PCI-X as I know I put SA6402's in several RZ2600's that I gave away when I moved to Texas. The ZX6000 should be indentical.
Post by Rod Regier
I suspect if it has an I64 processor and either a PCIx or PCIe buss
you could be in business.
Guess I'll cross off the Exeter vendor as a future option.
I'm hoping they send it next week. There was one positive feedback for the PCI-e to PCI card from a buyer that bought theirs late in March.
Post by Rod Regier
I've done some HW maint on the RX2660 but I didn't start out with
once as broken as yours.
I would be using the embedded ILO (System Event Log is your friend) and front panel lights to try and get a sense of what might be the lowest level issue.
hanging it. The iLO is a co-processor - it will work as long as the chassis has power.
Yes, the iLo is where I have been connecting. Last time I messed with it was in 2017 before moving and the last time it worked was in 2015.
--
John H. Reinhardt
Rod Regier
2020-04-18 19:42:03 UTC
Permalink
Got to the RX2600 + P410i + PLX8112 testing phase.

Worked fine, no issues with OpenVMS boot from SATA SSD
(external)

Was also able to run the firmware update util and have it recognize the P410i *thru* the PLX8112 adapter. Excellent.

DISKBLOCK benchmark data (no write ops) from past and current:

RX2600 1.6G SA6402 dual U320 15K drives 4G main memory

Total I/Os completed: 165613
Run time: 120 Seconds
I/Os per second: 1380.1
mS per I/O: 0.7
Throughput: 690.1 KBytes per second

RX2660 P400 SATA JBOD 500G 500 IOPS SSD

Total I/Os completed: 1017417
Run time: 120 Seconds
I/Os per second: 8478.5
mS per I/O: 0.1
Throughput: 4239.2 KBytes per second

RX2600 w/P410i no write cache

RX2600 1.6G P410i no write cache plus PLX8112 adapter Samsung SATA SSD 8G main memory

Total I/Os completed: 3812311
Run time: 120 Seconds
I/Os per second: 31769.3
mS per I/O: 0.0
Throughput: 15884.6 KBytes per second



Apparently the P410i is The Amazing Hulk compared to the SA6402 on the RX2600.

No slouch compared to a P400 on an RX2660 either
i***@dvdwrite.de
2020-04-18 19:57:57 UTC
Permalink
Post by Rod Regier
Got to the RX2600 + P410i + PLX8112 testing phase.
Worked fine, no issues with OpenVMS boot from SATA SSD
(external)
Was also able to run the firmware update util and have it recognize the P410i *thru* the PLX8112 adapter. Excellent.
RX2600 1.6G SA6402 dual U320 15K drives 4G main memory
Total I/Os completed: 165613
Run time: 120 Seconds
I/Os per second: 1380.1
mS per I/O: 0.7
Throughput: 690.1 KBytes per second
RX2660 P400 SATA JBOD 500G 500 IOPS SSD
Total I/Os completed: 1017417
Run time: 120 Seconds
I/Os per second: 8478.5
mS per I/O: 0.1
Throughput: 4239.2 KBytes per second
RX2600 w/P410i no write cache
RX2600 1.6G P410i no write cache plus PLX8112 adapter Samsung SATA SSD 8G main memory
Total I/Os completed: 3812311
Run time: 120 Seconds
I/Os per second: 31769.3
mS per I/O: 0.0
Throughput: 15884.6 KBytes per second
Apparently the P410i is The Amazing Hulk compared to the SA6402 on the RX2600.
No slouch compared to a P400 on an RX2660 either
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com
Great numbers!

How about using a cache (1 GB is possible). It might possible to reach
even higher numbers.

Eberhard
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
Rod Regier
2020-04-20 13:32:57 UTC
Permalink
re: write caching

My early milestone goal was bench minimum cost proof of function on RX2600.
Write cache support was an additional cost.

Now that the proof of function testing was successful, I'll look at acquiring FBWC capable cache memory along with the necessary super capacitor to enable that functionality. I'll then change the benchmark to use write operations and re-test.

I was happily surprised how good was the benchmark performance for the P410i.
(My primary goal was gaining access to SATA SSD support for the RX2600).
Rod Regier
2020-04-21 14:01:08 UTC
Permalink
On Sunday I initiated a 1000 pass UETP execution on my bench test system to evaluate stability of the configuration in a simulated production environment.
(I've used UETP many times in the past to identify intermittent system problems on Integrity and Alpha servers).

I've ordered 1G FBWC option for my base model P410i card from a US Ebay vendor.
I'll run suitable read *and* write cache benchmarks on my bench test configuration once that arrives in say 2 weeks. I would have been happy with the smaller FBWC option but the price on that 1G version was very attractive.
Rod Regier
2020-04-22 19:11:36 UTC
Permalink
Temporarily back onsite. My RX2600+P410i+PLX8112+SATA SSD UETP bench configuration burn was concluded by me manually at 735 passes. (Just a decent stopping point). That was around 93 hours elaqpsed.

So that looks pretty solid.

Now waiting on 1G FBWC option for the P410i to take testing into the next phase.

The PLX8112 adapter card PCIx buss interface would seem to be the data transfer bottleneck - PCI 32 bit 33MHz bus one slot.
This is mitigated somewhat by the P410i card onboard cache.

FWIW, the PCIx busses on the various topic system are described as:
RX2660 64 bit/133 MHz PCI
(or 266 depending on the I/O card cage selected)
RX2600 64 bit/133 MHz PCI
DS15 64 bit/66 MHz PCI best available
DS10 64 bit/66 MHz PCI best available
Ben Lambert
2020-04-22 21:06:01 UTC
Permalink
Post by Rod Regier
Temporarily back onsite. My RX2600+P410i+PLX8112+SATA SSD UETP bench configuration burn was concluded by me manually at 735 passes. (Just a decent stopping point). That was around 93 hours elaqpsed.
So that looks pretty solid.
Now waiting on 1G FBWC option for the P410i to take testing into the next phase.
The PLX8112 adapter card PCIx buss interface would seem to be the data transfer bottleneck - PCI 32 bit 33MHz bus one slot.
This is mitigated somewhat by the P410i card onboard cache.
RX2660 64 bit/133 MHz PCI
(or 266 depending on the I/O card cage selected)
RX2600 64 bit/133 MHz PCI
DS15 64 bit/66 MHz PCI best available
DS10 64 bit/66 MHz PCI best available
Enjoying reading your investigation/trials!

I may have missed it, but is there a reason you didn't go for the PLX8114 and get the benefits of the PCI-X interface instead of legacy PCI? The StarTech PCIX1PEX4 seems like a better option.

My DS20e is unfortunately hampered by the 5V-only slot keying, which I failed to notice until after I'd ordered a SA6404A, so I'm stuck with my SA5304A as the 'best' for the foreseeable future.
John H. Reinhardt
2020-04-22 21:20:38 UTC
Permalink
Post by Ben Lambert
Enjoying reading your investigation/trials!
I may have missed it, but is there a reason you didn't go for the PLX8114 and get the benefits of the PCI-X interface instead of legacy PCI? The StarTech PCIX1PEX4 seems like a better option.
I asked on Amazon about it but the StarTech only has a PIC-e 1X slot sized connector. You cannot put a 4x or 8x tabbed P410/P400 into it. Seems to be significant limitation. If you look at the pictures the PLX8112 cards have full x16 length slots even if they don't use them all.
--
John H. Reinhardt
Ben Lambert
2020-04-23 07:03:46 UTC
Permalink
Post by John H. Reinhardt
Post by Ben Lambert
Enjoying reading your investigation/trials!
I may have missed it, but is there a reason you didn't go for the PLX8114 and get the benefits of the PCI-X interface instead of legacy PCI? The StarTech PCIX1PEX4 seems like a better option.
I asked on Amazon about it but the StarTech only has a PIC-e 1X slot sized connector. You cannot put a 4x or 8x tabbed P410/P400 into it. Seems to be significant limitation. If you look at the pictures the PLX8112 cards have full x16 length slots even if they don't use them all.
--
John H. Reinhardt
That's the PCIX1PEX1. The PCIX1PEX4, has a PCI-e 16x connector, and a PCI-e 4x interface. Although stock seems scarce at the moment.

https://www.startech.com/eu/support/PCIX1PEX4
https://www.newegg.com/startech-com-model-pcix1pex4-pci-to-pci-express/p/N82E16815158315
https://www.amazon.com/StarTech-PCI-X-Express-Adapter-PCIX1PEX4/dp/B004H603SW/ref=cm_cr_arp_d_product_top?ie=UTF8

Ben.
John H. Reinhardt
2020-04-23 07:39:07 UTC
Permalink
Post by Ben Lambert
Post by John H. Reinhardt
Post by Ben Lambert
Enjoying reading your investigation/trials!
I may have missed it, but is there a reason you didn't go for the PLX8114 and get the benefits of the PCI-X interface instead of legacy PCI? The StarTech PCIX1PEX4 seems like a better option.
I asked on Amazon about it but the StarTech only has a PIC-e 1X slot sized connector. You cannot put a 4x or 8x tabbed P410/P400 into it. Seems to be significant limitation. If you look at the pictures the PLX8112 cards have full x16 length slots even if they don't use them all.
--
John H. Reinhardt
That's the PCIX1PEX1. The PCIX1PEX4, has a PCI-e 16x connector, and a PCI-e 4x interface. Although stock seems scarce at the moment.
https://www.startech.com/eu/support/PCIX1PEX4
https://www.newegg.com/startech-com-model-pcix1pex4-pci-to-pci-express/p/N82E16815158315
https://www.amazon.com/StarTech-PCI-X-Express-Adapter-PCIX1PEX4/dp/B004H603SW/ref=cm_cr_arp_d_product_top?ie=UTF8
Ben.
Ah. I see. That would be the one to get. However, if it's out of stock everywhere then there really is no choice.
--
John H. Reinhardt
Rod Regier
2020-04-22 19:36:03 UTC
Permalink
Anticipating needing a possible future onsite spare for the PLX8112 adapter card, I have ordered one from an AliExpress vendor. Since the card is actually made in China, that is not as outlandish as it sounds.

https://www.aliexpress.com/item/4000283624785.html?spm=a2g0o.productlist.0.0.8ab549d20mMHj5

998 pieces in stock - US $23.98 ea, currently on 35% discount.

Shipping: US $2.21
to Canada via AliExpress Standard Shipping
Estimated Delivery: 15-30 days

If I was in a bigger hurry:

Shipping: US $31.64
to Canada via DHL
Estimated Delivery: 7-16 days

All of the US and European sellers buy batches from China and resell them.
Rod Regier
2020-04-23 14:15:28 UTC
Permalink
I'm not 100% certain, but my research suggests that the PCIX1PEX4 card is currently out of production with no stock anywhere except maybe in some obscure retail corners.
Rod Regier
2020-04-23 14:26:29 UTC
Permalink
Hi Ben

I went with the PLX8112 card to build on Eberhard's testing success rather than re-certify a different adapter for my initial testing.

see this thread:

https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw

My foremost goal is to support current SATA SSD's on RX2600 as bootable RAID set with performance at least comparable to the SA6402 using parallel SCSI 15KRPM disks also in a RAID set.

If I could get *better* performance that would be a happy bonus.

My past experience with other Startech products that needed separate software driver support has been less than stellar, so they would not be my first choice.
It they worked without needing a Startech driver that would be better.

Of course, there would be less than zero prospect of there being any Startech OpenVMS/I64 driver support :-)

If I could find any of the unobtanium PCIX1PEX4 cards I might explore those in the future. That would eliminate the most serious transfer bottleneck. The quest for performance never ends :-)
Eberhard Heuser
2020-04-23 15:36:18 UTC
Permalink
Post by Rod Regier
Hi Ben
I went with the PLX8112 card to build on Eberhard's testing success rather than re-certify a different adapter for my initial testing.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
My foremost goal is to support current SATA SSD's on RX2600 as bootable RAID set with performance at least comparable to the SA6402 using parallel SCSI 15KRPM disks also in a RAID set.
If I could get *better* performance that would be a happy bonus.
My past experience with other Startech products that needed separate software driver support has been less than stellar, so they would not be my first choice.
It they worked without needing a Startech driver that would be better.
Of course, there would be less than zero prospect of there being any Startech OpenVMS/I64 driver support :-)
If I could find any of the unobtanium PCIX1PEX4 cards I might explore those in the future. That would eliminate the most serious transfer bottleneck. The quest for performance never ends :-)
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com
A PCI2PCIe "converter" needs no driver because it is a passive
component. But my thoughts may be wrong.

I have used Acard SCSI2SATA converters and they need no driver, too.

Eberhard
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
John H. Reinhardt
2020-04-28 16:51:05 UTC
Permalink
A PCI2PCIe "converter" needs no driver because it is a passive component. But my thoughts may be wrong.
I would agree with that. Any "driver" would be in the internals in the bridge chip.
I have used Acard SCSI2SATA converters and they need no driver, too.
Eberhard
--
John H. Reinhardt
John H. Reinhardt
2020-04-28 16:49:17 UTC
Permalink
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter plus cabling).  The SATA drives will be deployed in an external enclosure.  I have a bill of materials drafted.  Parts are trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires 3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the Technical Summary documents on both, it's a PCI V2.1 compliant bus supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3 compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 - <https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a P400 or P800 series SATA HBA?
The PCI-e to PCI bridge board arrived yesterday. <ttps://www.ebay.com/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013>

The PCI tab fit into the DS10 PCI slot with out a problem. I had to take the bracket off the P410 as even the short one wouldn't fit correctly. I then wire tied it to the bracket on the PCI adapter card to keep it from drooping (the PCI card cage on a DS10 sits on it's side compared to most). I put it in the bottom PCI slot on the DS10 which is the 32-bit slot.

I did a quick power up test and, as expected, the P410 shows up in the config listing, but it's not bootable.

PCI Hose 00
Bus 00 Slot 07: Acer Labs M1543C
Bridge to Bus 1, ISA
Bus 00 Slot 09: DE500-BA Network Controller
ewa0.0.0.9.0 00-10-64-30-25-AB
Bus 00 Slot 11: DE500-BA Network Controller
ewb0.0.0.11.0 00-10-64-30-25-A9
Bus 00 Slot 13: Acer Labs M1543C IDE
dqa.0.0.13.0
Bus 00 Slot 14: 811110B5
Bridge to Bus 2, PCI
Bus 00 Slot 16: QLogic ISP10x0
pka0.7.0.16.0 SCSI Bus ID 7
Bus 00 Slot 17: Intel 21154-*E
Bridge to Bus 3, PCI
Bus 02 Slot 00: 323A103C/3243103C

Bus 03 Slot 04: FCA-2354
pga0.0.0.3004.0 WWN 1000-0000-c943-8
Bus 03 Slot 05: FCA-2354
pgb0.0.0.3005.0 WWN 1000-0000-c943-8



The P400 shows up at Bus 02, Slot00. PCI Id 3243103C (or word swapped 103C3243)

So now I need to put the rest of the system back together and boot OpenVMS and see if I can get MSA$UTIL to recognise it.
--
John H. Reinhardt
Eberhard Heuser
2020-04-28 17:56:35 UTC
Permalink
If someone provides the source code of the srm-console it would be possible to boot the alpha just like itaniums can do.
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form
factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units
w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter
plus cabling).  The SATA drives will be deployed in an external
enclosure.  I have a bill of materials drafted.  Parts are trickling in
for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires
3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the
Technical Summary documents on both, it's a PCI V2.1 compliant bus
supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3
compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 -
<https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a
P400 or P800 series SATA HBA?
The PCI-e to PCI bridge board arrived yesterday.
<ttps://www.ebay.com/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013>
The PCI tab fit into the DS10 PCI slot with out a problem. I had to
take the bracket off the P410 as even the short one wouldn't fit
correctly. I then wire tied it to the bracket on the PCI adapter card
to keep it from drooping (the PCI card cage on a DS10 sits on it's side
compared to most). I put it in the bottom PCI slot on the DS10 which
is the 32-bit slot.
I did a quick power up test and, as expected, the P410 shows up in the
config listing, but it's not bootable.
PCI Hose 00
Bus 00 Slot 07: Acer Labs M1543C
Bridge to Bus 1, ISA
Bus 00 Slot 09: DE500-BA Network Controller
ewa0.0.0.9.0 00-10-64-30-25-AB
Bus 00 Slot 11: DE500-BA Network Controller
ewb0.0.0.11.0 00-10-64-30-25-A9
Bus 00 Slot 13: Acer Labs M1543C IDE
dqa.0.0.13.0
Bus 00 Slot 14: 811110B5
Bridge to Bus 2, PCI
Bus 00 Slot 16: QLogic ISP10x0
pka0.7.0.16.0 SCSI Bus ID 7
Bus 00 Slot 17: Intel 21154-*E
Bridge to Bus 3, PCI
Bus 02 Slot 00: 323A103C/3243103C
Bus 03 Slot 04: FCA-2354
pga0.0.0.3004.0 WWN 1000-0000-c943-8
Bus 03 Slot 05: FCA-2354
pgb0.0.0.3005.0 WWN 1000-0000-c943-8
The P400 shows up at Bus 02, Slot00. PCI Id 3243103C (or word swapped 103C3243)
So now I need to put the rest of the system back together and boot
OpenVMS and see if I can get MSA$UTIL to recognise it.
--
John H. Reinhardt
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com
--
Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet.
John H. Reinhardt
2020-04-28 18:20:44 UTC
Permalink
Post by Eberhard Heuser
If someone provides the source code of the srm-console it would be possible to boot the alpha just like itaniums can do.
That would be interesting. As far as I know the SRM source never made it out of DEC/Compaq/HP unfortunately.
Post by Eberhard Heuser
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form
factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units
w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter
plus cabling).  The SATA drives will be deployed in an external
enclosure.  I have a bill of materials drafted.  Parts are trickling in
for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires
3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the
Technical Summary documents on both, it's a PCI V2.1 compliant bus
supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3
compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 -
<https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a
P400 or P800 series SATA HBA?
The PCI-e to PCI bridge board arrived yesterday.
<ttps://www.ebay.com/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013>
The PCI tab fit into the DS10 PCI slot with out a problem. I had to
take the bracket off the P410 as even the short one wouldn't fit
correctly. I then wire tied it to the bracket on the PCI adapter card
to keep it from drooping (the PCI card cage on a DS10 sits on it's side
compared to most). I put it in the bottom PCI slot on the DS10 which
is the 32-bit slot.
I did a quick power up test and, as expected, the P410 shows up in the
config listing, but it's not bootable.
PCI Hose 00
Bus 00 Slot 07: Acer Labs M1543C
Bridge to Bus 1, ISA
Bus 00 Slot 09: DE500-BA Network Controller
ewa0.0.0.9.0 00-10-64-30-25-AB
Bus 00 Slot 11: DE500-BA Network Controller
ewb0.0.0.11.0 00-10-64-30-25-A9
Bus 00 Slot 13: Acer Labs M1543C IDE
dqa.0.0.13.0
Bus 00 Slot 14: 811110B5
Bridge to Bus 2, PCI
Bus 00 Slot 16: QLogic ISP10x0
pka0.7.0.16.0 SCSI Bus ID 7
Bus 00 Slot 17: Intel 21154-*E
Bridge to Bus 3, PCI
Bus 02 Slot 00: 323A103C/3243103C
Bus 03 Slot 04: FCA-2354
pga0.0.0.3004.0 WWN 1000-0000-c943-8
Bus 03 Slot 05: FCA-2354
pgb0.0.0.3005.0 WWN 1000-0000-c943-8
The P400 shows up at Bus 02, Slot00. PCI Id 3243103C (or word swapped 103C3243)
So now I need to put the rest of the system back together and boot
OpenVMS and see if I can get MSA$UTIL to recognise it.
--
John H. Reinhardt
Rod Regier
2020-04-28 19:42:48 UTC
Permalink
John:

re: P410, P400 boot support

As Eberhard had mentioned earlier, the boot media itself will need to contain a version of:

sys$system:SYS$USER_CONFIG.DAT:

device = "Smart Array P-series P400"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x3230103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device
device = "Smart Array P-series PMC PCIe"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x323A103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device
John H. Reinhardt
2020-04-30 05:41:22 UTC
Permalink
Post by Rod Regier
re: P410, P400 boot support
Yep. Got that. I think I need a newer OpenVMS. Tried it on an OpenVMS Alpha V8.3 and got this result below. Either that or maybe newer firmware in the P410. It's a6.00. I don't know how to upgrade it though. I have nothing else it will work in. Hopefully V8.4 will work.

$ type sys$system:SYS$USER_CONFIG.DAT

device = "Smart Array P-series P400"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x3230103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device

device = "Smart Array P-series PMC PCIe"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x323A103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device

$ show dev pk

Device Device Error
Name Status Count
PKA0: Online 1
PKB0: Online 0
PKC0: Online 0

$ show dev/full pka0

Device PKA0:, device type HP Smart Array, is online, error logging is enabled.

Error count 1 Operations completed 4
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G,W
Reference count 0 Default buffer size 65535
Current preferred CPU Id 0 Fastpath 1
Current Interrupt CPU Id 0


$ mcr msa$util
MSA> help

Information available:

ACCEPT ADD DELETE EXIT FLASH HELP HPSAS
LOCATE MSA Overview READ RESET SCAN SET
SHOW SmartArray START

Topic?MSA> set controller PKA0:
MSA> show controller

Adapter: _PKA0: (DEFAULT)
P410 (c) HP PACCRID125002W7 Software 6.00
SCSI_VERSION = X3.131:1994 (SCSI-2)
Supported Redundancy Mode:
Not currently Redundant
Current Role: Active
Cache:
912 megabyte read/write cache
Cache is not configured, and Cache is disabled.
No unflushed data in cache.
Battery:
Battery is fully charged.

MSA> show disks
Error getting redundant controller parameters from controller.
%SYSTEM-F-CTRLERR, fatal controller error
$
--
John H. Reinhardt
i***@dvdwrite.de
2020-04-30 07:38:10 UTC
Permalink
Post by John H. Reinhardt
Software 6.00
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e

Try to update the firmware.

PC-only software, sorry.

Eberhard
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
John H. Reinhardt
2020-04-30 14:59:17 UTC
Permalink
Post by i***@dvdwrite.de
Post by John H. Reinhardt
Software 6.00
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e
Try to update the firmware.
PC-only software, sorry.
Eberhard
Yes. That's a dilemma. I just realized I do have the ZX6000 Itanium machine. If there is an IA64 firmware updater then I might be able using the same PCI Bridge card.
--
John H. Reinhardt
i***@dvdwrite.de
2020-04-30 16:09:54 UTC
Permalink
Post by i***@dvdwrite.de
Post by John H. Reinhardt
Software 6.00
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e
Try to update the firmware.
PC-only software, sorry.
Eberhard
Yes. That's a dilemma.  I just realized I do have the ZX6000 Itanium
machine. If there is an IA64 firmware updater then I might be able
using the same PCI Bridge card.
https://support.hpe.com/hpsc/swd/public/detail?swItemId=ux-70922-1
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
i***@dvdwrite.de
2020-04-30 16:07:05 UTC
Permalink
Post by i***@dvdwrite.de
Post by John H. Reinhardt
Software 6.00
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e
Try to update the firmware.
PC-only software, sorry.
Eberhard
Yes. That's a dilemma.  I just realized I do have the ZX6000 Itanium
machine. If there is an IA64 firmware updater then I might be able
using the same PCI Bridge card.
saupdate.efi can do this.
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
i***@dvdwrite.de
2020-04-30 16:15:51 UTC
Permalink
Post by i***@dvdwrite.de
Post by John H. Reinhardt
Software 6.00
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e
Try to update the firmware.
PC-only software, sorry.
Eberhard
Yes. That's a dilemma.  I just realized I do have the ZX6000 Itanium
machine. If there is an IA64 firmware updater then I might be able
using the same PCI Bridge card.
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e
download exe-file.
extract firmware.bin by unpacking the exe-file with 7-zip on a pc.

Copy firmare.bin and saupdate.efi on a USB-Stick. Start the
Itanium-machine and switch to efi.

start saupdate.efi.

You'll get instuctions how to update a controller.

Good luck
Eberhard
--
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus
John H. Reinhardt
2020-05-02 16:22:50 UTC
Permalink
Post by i***@dvdwrite.de
Post by John H. Reinhardt
Software 6.00
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e
Try to update the firmware.
PC-only software, sorry.
Eberhard
Yes. That's a dilemma.  I just realized I do have the ZX6000 Itanium machine. If there is an IA64 firmware updater then I might be able using the same PCI Bridge card.
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-993349aa061b4897b5ae4d212e > download exe-file.
extract firmware.bin by unpacking the exe-file with 7-zip on a pc.
Copy firmare.bin and saupdate.efi on a USB-Stick. Start the Itanium-machine and switch to efi.
start saupdate.efi.
You'll get instuctions how to update a controller.
Good luck
Eberhard
Thanks Eberhard. Rod has also supplied a link for the Itanium version of the firmware so I can use that as well.
--
John H. Reinhardt
John E. Malmberg
2020-05-01 22:55:51 UTC
Permalink
Post by Rod Regier
re: P410, P400 boot support
Have you tried:

P0>>> set HEAP_EXPAND 2MB
P0>>> initialize
P0>>> set BOOTBIOS pya0
P0>>> initialize

Regards,
-John
John H. Reinhardt
2020-05-02 16:15:42 UTC
Permalink
Post by John E. Malmberg
Post by Rod Regier
re: P410, P400 boot support
P0>>> set HEAP_EXPAND 2MB
P0>>> initialize
P0>>> set BOOTBIOS pya0
P0>>> initialize
Regards,
-John
Thanks John. Unfortunately I cannot do that as the SRM console does not recognize the P140 as a valid device, therefor it gets no device name such as PYA0.

This is what the SRM conosle lists for "SHOW CONFIG"

PCI Hose 00
Bus 00 Slot 07: Acer Labs M1543C
Bridge to Bus 1, ISA
Bus 00 Slot 09: DE500-BA Network Controller
ewa0.0.0.9.0 00-10-64-30-25-AB
Bus 00 Slot 11: DE500-BA Network Controller
ewb0.0.0.11.0 00-10-64-30-25-A9
Bus 00 Slot 13: Acer Labs M1543C IDE
dqa.0.0.13.0
Bus 00 Slot 14: 811110B5
Bridge to Bus 2, PCI
Bus 00 Slot 16: QLogic ISP10x0
pka0.7.0.16.0 SCSI Bus ID 7
Bus 00 Slot 17: Intel 21154-*E
Bridge to Bus 3, PCI
Bus 02 Slot 00: 323A103C/3243103C

Bus 03 Slot 04: FCA-2354
pga0.0.0.3004.0 WWN 1000-0000-c943-8
Bus 03 Slot 05: FCA-2354
pgb0.0.0.3005.0 WWN 1000-0000-c943-8


The P410 is at Bus 02, Slot 00. The SRM does not recognize the PCI identification code of "323A103C/3243103C" so it does not create a device name. Which means I cannot run the BIOS as far as I know.
--
John H. Reinhardt
Rod Regier
2020-04-30 17:19:48 UTC
Permalink
I've successfully run the P410 I64 firmware updater thru EFI on an RX2660 and an RX2600 using the adapter card.

Highest rev of the P410 firmware I could find was 6.64(5 Oct 2017)

This was where I found my copy:

https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-a17b1b7dcc284cf5a21ecfa7c6#tab1
John H. Reinhardt
2020-05-02 16:22:07 UTC
Permalink
Post by Rod Regier
I've successfully run the P410 I64 firmware updater thru EFI on an RX2660 and an RX2600 using the adapter card.
Highest rev of the P410 firmware I could find was 6.64(5 Oct 2017)
https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-a17b1b7dcc284cf5a21ecfa7c6#tab1
Thanks. I have downloaded that. Now I just have to get my ZX6000 up and running. It's been sitting since my move to Texas 2 years ago.
--
John H. Reinhardt
John H. Reinhardt
2020-05-03 21:28:20 UTC
Permalink
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter plus cabling).  The SATA drives will be deployed in an external enclosure.  I have a bill of materials drafted.  Parts are trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires 3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the Technical Summary documents on both, it's a PCI V2.1 compliant bus supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3 compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 - <https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a P400 or P800 series SATA HBA?
Some limited success with OpenVMS Alpha V8.4. After updating the SYS$SYSTEM:SYS$USER_CONFIG.DAT with:

device = "Smart Array P-series P400"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x3230103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device

device = "Smart Array P-series PMC PCIe"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x323A103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device

And rebooting, I now have:

$ show dev p

Device Device Error
Name Status Count
PKB0: Online 0
PKA0: Online 0
PGA0: Online 0
PGB0: Online 0
PKC0: Online 0


PKA0: is the Smart Array P410. PKB0: and PKC0: are the two channels for the 3X-KZPEA Adaptec SCSI.

$ show dev/full pka0:

Device PKA0:, device type HP Smart Array, is online, error logging is enabled.

Error count 0 Operations completed 2854
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G,W
Reference count 0 Default buffer size 65535
Current preferred CPU Id 0 Fastpath 1
Current Interrupt CPU Id 0

Still no disks showing up, of course.

$ show dev d

Device Device Error Volume Free Trans Mnt
Name Status Count Label Blocks Count Cnt
ASIMOV$DKB0: Mounted 0 ASIMOV084 14419878 313 1
ASIMOV$DKB100: Online 0
ASIMOV$DVA0: Online 0
ASIMOV$DQA0: Online 0
ASIMOV$DQA1: Offline 1
ASIMOV$DQB0: Offline 1
ASIMOV$DQB1: Offline 1
ASIMOV$DKC0: Online 0


So now when I run MSA$UTIL things work better.

$ mcr MSA$UTIL

MSA> show controller

Adapter: _PKA0: (DEFAULT)
P410 (c) HP PACCRID125002W7 Software 6.00
Port Address: 50014380-22aa92a0
Supported Redundancy Mode: Not Available.
Cache:
912 megabyte read/write cache
Cache is not configured, and Cache is disabled.
No unflushed data in cache.
Battery:
Battery is fully charged.
MSA>

My two Samsung 860EVO 256GB SSD disk show up as expected in bays 5 & 6 in the 6-bay SATA enclosure.

MSA> show disks

SATA device [Disk]
Disk 25505: bus: 1, Target id: 4, Port: 2I, Box/Enclosure: 255, Bay: 5, size 232.89 [250.06]GB
Disk 25505, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.

SATA device [Disk]
Disk 25506: bus: 1, Target id: 5, Port: 2I, Box/Enclosure: 255, Bay: 6, size 232.89 [250.06]GB
Disk 25506, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.

Here's where things go bad again. Trying to make units of the disks.

MSA> add unit 1000/disk=25505/jbod/partition=0
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505/jbod/partition=0/size=32gb
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505/jbod
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505/raid=1
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505
Specified disk is not a Valid disk
MSA> add unit 1000/disk=2550
Specified disk is not a Valid disk
MSA> add unit 1000/disk=5
The selected controller does not support this unit number only supports 64 units
MSA> add unit 10/disk=5/jbod
A non-existant/invalid partition was specified.
Partition numbers for units on a disk should be sequential.
Please create units with appropriate partition numbers.
MSA> add unit 10/disk=5/jbod/partition=0
A non-existant/invalid partition was specified.
Partition numbers for units on a disk should be sequential.
Please create units with appropriate partition numbers.
MSA> show disk 5
Disk 5 does not exist
MSA> show disk 255
Disk 255 does not exist
MSA> show disk 2555
Specified disk is not a Valid disk
MSA> show disk 25505
Specified disk is not a Valid disk
MSA> show disks

SATA device [Disk]
Disk 25505: bus: 1, Target id: 4, Port: 2I, Box/Enclosure: 255, Bay: 5, size 232.89 [250.06]GB
Disk 25505, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.

SATA device [Disk]
Disk 25506: bus: 1, Target id: 5, Port: 2I, Box/Enclosure: 255, Bay: 6, size 232.89 [250.06]GB
Disk 25506, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.
MSA>


Anyone have any ideas? Next on my list is to try to update the firmware to 6.65 but I didn't see anything in the release notes that might indicate a fix for this.

Also, the MSA$UTIL has a firmware loading option. Anyone used it?

FLASH

FIRMWARE


FLASH FIRMWARE is used to update the firmware of a given
controller. Specify the firmware filename to be used.

Format:
FLASH FIRMWARE <file name> / <qualifiers>

Additional information available:

Parameters Qualifiers
/VERBOSE
Restrictions
--
John H. Reinhardt
Eberhard Heuser
2020-05-04 06:35:51 UTC
Permalink
What you can do ist to configure the Controller with your itanium machine by using EFI and move the equipment back to your Alpha.
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form
factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units
w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter
plus cabling).  The SATA drives will be deployed in an external
enclosure.  I have a bill of materials drafted.  Parts are trickling in
for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires
3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the
Technical Summary documents on both, it's a PCI V2.1 compliant bus
supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3
compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 -
<https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a
P400 or P800 series SATA HBA?
Some limited success with OpenVMS Alpha V8.4. After updating the
device = "Smart Array P-series P400"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x3230103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device
device = "Smart Array P-series PMC PCIe"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x323A103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT
end_device
$ show dev p
Device Device Error
Name Status Count
PKB0: Online 0
PKA0: Online 0
PGA0: Online 0
PGB0: Online 0
PKC0: Online 0
PKA0: is the Smart Array P410. PKB0: and PKC0: are the two channels for
the 3X-KZPEA Adaptec SCSI.
Device PKA0:, device type HP Smart Array, is online, error logging is enabled.
Error count 0 Operations completed
2854
Owner process "" Owner UIC
[SYSTEM]
Owner process ID 00000000 Dev Prot
S:RWPL,O:RWPL,G,W
Reference count 0 Default buffer size
65535
Current preferred CPU Id 0 Fastpath
1
Current Interrupt CPU Id 0
Still no disks showing up, of course.
$ show dev d
Device Device Error Volume Free
Trans Mnt
Name Status Count Label Blocks Count Cnt
ASIMOV$DKB0: Mounted 0 ASIMOV084 14419878
313 1
ASIMOV$DKB100: Online 0
ASIMOV$DVA0: Online 0
ASIMOV$DQA0: Online 0
ASIMOV$DQA1: Offline 1
ASIMOV$DQB0: Offline 1
ASIMOV$DQB1: Offline 1
ASIMOV$DKC0: Online 0
So now when I run MSA$UTIL things work better.
$ mcr MSA$UTIL
MSA> show controller
Adapter: _PKA0: (DEFAULT)
P410 (c) HP PACCRID125002W7 Software 6.00
Port Address: 50014380-22aa92a0
Supported Redundancy Mode: Not Available.
912 megabyte read/write cache
Cache is not configured, and Cache is disabled.
No unflushed data in cache.
Battery is fully charged.
MSA>
My two Samsung 860EVO 256GB SSD disk show up as expected in bays 5 & 6
in the 6-bay SATA enclosure.
MSA> show disks
SATA device [Disk]
5, size 232.89 [250.06]GB
Disk 25505, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.
SATA device [Disk]
6, size 232.89 [250.06]GB
Disk 25506, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.
Here's where things go bad again. Trying to make units of the disks.
MSA> add unit 1000/disk=25505/jbod/partition=0
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505/jbod/partition=0/size=32gb
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505/jbod
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505/raid=1
Specified disk is not a Valid disk
MSA> add unit 1000/disk=25505
Specified disk is not a Valid disk
MSA> add unit 1000/disk=2550
Specified disk is not a Valid disk
MSA> add unit 1000/disk=5
The selected controller does not support this unit number only supports 64 units
MSA> add unit 10/disk=5/jbod
A non-existant/invalid partition was specified.
Partition numbers for units on a disk should be sequential.
Please create units with appropriate partition numbers.
MSA> add unit 10/disk=5/jbod/partition=0
A non-existant/invalid partition was specified.
Partition numbers for units on a disk should be sequential.
Please create units with appropriate partition numbers.
MSA> show disk 5
Disk 5 does not exist
MSA> show disk 255
Disk 255 does not exist
MSA> show disk 2555
Specified disk is not a Valid disk
MSA> show disk 25505
Specified disk is not a Valid disk
MSA> show disks
SATA device [Disk]
5, size 232.89 [250.06]GB
Disk 25505, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.
SATA device [Disk]
6, size 232.89 [250.06]GB
Disk 25506, # 0, size 488331632 blocks, (232.85 [250.03] GB), Unused.
MSA>
Anyone have any ideas? Next on my list is to try to update the
firmware to 6.65 but I didn't see anything in the release notes that
might indicate a fix for this.
Also, the MSA$UTIL has a firmware loading option. Anyone used it?
FLASH
FIRMWARE
FLASH FIRMWARE is used to update the firmware of a given
controller. Specify the firmware filename to be used.
FLASH FIRMWARE <file name> / <qualifiers>
Parameters Qualifiers
/VERBOSE
Restrictions
--
John H. Reinhardt
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com
--
Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet.
John H. Reinhardt
2020-05-26 14:54:45 UTC
Permalink
Post by Eberhard Heuser
What you can do ist to configure the Controller with your itanium machine by using EFI and move the equipment back to your Alpha.
Post by John H. Reinhardt
Anyone have any ideas? Next on my list is to try to update the
firmware to 6.65 but I didn't see anything in the release notes that
might indicate a fix for this.
Also, the MSA$UTIL has a firmware loading option. Anyone used it?
FLASH
FIRMWARE
FLASH FIRMWARE is used to update the firmware of a given
controller. Specify the firmware filename to be used.
FLASH FIRMWARE <file name> / <qualifiers>
Parameters Qualifiers
/VERBOSE
Restrictions
--
John H. Reinhardt
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com
I finally had time to try the MSA$UTIL FLASH FIRMWARE command but no luck. Using the /VERBOSE qualifier I get one line of output and then the process goes into RWAST and hangs. The only recourse is to reboot.

$ dir /date/size

Directory DKB0:[KITS.STORAGE.BL8x0ci2IOSASFW664_manual]

fwupdate.nsh;1 2 25-MAY-2020 14:38:12.95
sandman664b0.bin;1 8192 25-MAY-2020 14:38:12.83
saupdate.efi;1 883 25-MAY-2020 14:38:13.26

Total of 3 files, 9077 blocks.
$ mcr msa$util
MSA> set controller pka
MSA> show controller

Adapter: _PKA0: (DEFAULT)
P410 (c) HP PACCRID125002W7 Software 6.00
Port Address: 50014380-22aa92a0
Supported Redundancy Mode: Not Available.
Cache:
912 megabyte read/write cache
Cache is not configured, and Cache is disabled.
No unflushed data in cache.
Battery:
Battery is fully charged.
MSA> flash firmware /verbose sandman664b0.bin
Loading chunk 1, read = 1 chunks of 16384 size
Interrupt


Looks like I have to dig out the ZX6000 and try an IA64 EFI firmware update.
--
John H. Reinhardt
John H. Reinhardt
2020-08-08 23:15:11 UTC
Permalink
This post might be inappropriate. Click to display it.
Rod Regier
2020-05-04 13:20:38 UTC
Permalink
You definitely want to update the P410 firmware to current before trying to proceeding any further. Not all bugs are documented in the firmware release notes.

I've never seen 5 digit unit numbers for any of the MSA family controllers so that seems very suspicious too.

As Eberhard mentioned, you can configure the controller and attached disks on another box and them move them over. Just remember you'll need the SYS$SYSTEM:SYS$USER_CONFIG.DAT addition and a reboot on any box you want to access the P410 controller with OpenVMS.
Rod Regier
2020-05-08 17:13:43 UTC
Permalink
DISKBENCH execution Fixed details:

Copy of 28800 block MACRO.EXE as test file
80:20 read/write ratio
120 second test duration
I/O per second
Thruput is in units of KBy/sec

IOPS Thruput MSperIO Model cache detail MS detail Server Buss Notes
14086 7043 0.1 SA6402 64MR+64MW Mag 2*15K 80P RX2600 PCIX DYMIT
40950 20475 0.0 P410 228MR+684MW SSD RX2600 PCIe+PCIX adapt bench
49279 24639 0.0 P410 228MR+684MW SSD RX2660 PCIe bench
50949 25474 0.0 P400 52MR+145MW SSD RX2660 PCIe bench
57646 28823 0.0 P400 52MR+156MW 2*SSD RX2660 PCIe P217
David Turner
2020-06-25 17:30:38 UTC
Permalink
Anyone thought of using USB to connect to external SSD
I use cineraid 2 bay USB 3.0 on my computer here with 2 sata ssds
fast and easy to hook up
I imagine there are at least some USB 3.0 cards that could be worked in
with the rx2660 or use the onboard one.
Post by Rod Regier
Copy of 28800 block MACRO.EXE as test file
80:20 read/write ratio
120 second test duration
I/O per second
Thruput is in units of KBy/sec
IOPS Thruput MSperIO Model cache detail MS detail Server Buss Notes
14086 7043 0.1 SA6402 64MR+64MW Mag 2*15K 80P RX2600 PCIX DYMIT
40950 20475 0.0 P410 228MR+684MW SSD RX2600 PCIe+PCIX adapt bench
49279 24639 0.0 P410 228MR+684MW SSD RX2660 PCIe bench
50949 25474 0.0 P400 52MR+145MW SSD RX2660 PCIe bench
57646 28823 0.0 P400 52MR+156MW 2*SSD RX2660 PCIe P217
Stephen Hoffman
2020-06-25 18:33:11 UTC
Permalink
Post by David Turner
Anyone thought of using USB to connect to external SSD
OpenVMS USB support was USB 2.0 (Low Speed, Full Speed, and/or High
Speed) when last I checked, and none of those USB speeds will make
particularly effective use of an SSD.
USB High Speed is rated at 480 Mbps. That's approximately FireWire 400 speeds.
You'd want USB 3 (SuperSpeed USB 5Gbps, SuperSpeed USB 10Gbps, or
SuperSpeed USB 20Gbps, or USB-C with Thunderbolt 3, or USB4, or
Thunderbolt) for an SSD.
No currently-supported OpenVMS servers offer that USB hardware support.
VSI will undoubtedly be looking at USB SuperSpeed and USB-C and its
permutations, but that support will likely only become available on
x86-64.
--
Pure Personal Opinion | HoffmanLabs LLC
John E. Malmberg
2020-06-26 12:47:24 UTC
Permalink
Post by Stephen Hoffman
Post by David Turner
Anyone thought of using USB to connect to external SSD
OpenVMS USB support was USB 2.0 (Low Speed, Full Speed, and/or High
Speed) when last I checked, and none of those USB speeds will make
particularly effective use of an SSD.
Did USB 2.0 support ever make it into Alpha?

I seem to recall it only listed for IA64.

A network booted or CD-booted Alpha/IA64 can use USB disks up to the
size limit supported by OpenVMS.

Regards,
-John
Stephen Hoffman
2020-06-26 16:21:10 UTC
Permalink
Post by John E. Malmberg
Post by Stephen Hoffman
Post by David Turner
Anyone thought of using USB to connect to external SSD
OpenVMS USB support was USB 2.0 (Low Speed, Full Speed, and/or High
Speed) when last I checked, and none of those USB speeds will make
particularly effective use of an SSD.
Did USB 2.0 support ever make it into Alpha?
I seem to recall it only listed for IA64.
TL;DR: SSD connected to OpenVMS on something other than PCIe SAS/SATA
or Fibre Channel... isn't that beneficial, performance-wise. If it
works.


Any discussions of the dearth of feature and support documentation
aside, and that all eventual fodder for future VSI doc updates...


Official (AFAIK) USB support...

On Alpha, specific keyboard and mouse human interface devices and only
on EV7 Marvel-class boxes, via OHCI (slow).

On Itanium, support includes specific human interface devices, specific
optical devices, and specific terminal muxes. Via OHCI and EHCI, IIRC.

No, I don't know which specific USB devices are supported, nor do I
know of any comprehensive listings of supported widgets from either HPE
nor VSI. What's available is seemingly scattered all over.

Some Itanium boxes have or were supposed to have a USB-connected
optical drive. They'll show as DN devices.


As differentiated from official support...

USB 2.0 EHCI (less slow) reportedly does mostly-work on OpenVMS Alpha
V8.3 and later.

This where the USB controller itself works, which on Alpha is probably
only Marvel, or an add-on PCI USB adapter.

Some of the server-integrated USB controllers do not work.

If the controller works, other USB devices can work, and variously do.

From working with them, some flash drives work with OpenVMS, and some
don't. How to reliably distinguish those which will from those which
won't?

AlphaServer DS10 and AlphaServer DS10L integrated USB didn't work, and
is not now and likely won't ever be supported, so you'd need a
third-party USB PCI adapter for that, and some of those add-ons do work.

IIRC, there's some info on the USB PCI adapters with the USB chipsets
that work posted around, though approximately no USB PCI adapter will
list that chipset on the box.

PCI USB adapters are probably still scrounge-able.

I would doubt VSI will spend a great deal of effort on adding to the
existing USB support and adding USB-C support (USB4 does add
Thunderbolt, etc), as most of that tends to be client-side past
(potentially transiently-connected) keyboard, mouse and optical. And
that direct-connected HID usage is likely generally declining. At best,
we'll probably see better flash support and HDD support added. Maybe.

A USB SDA extension exists and can be useful for poking around.

The HPE-era USB code was common across Alpha and Itanium, save for a
few Marvel-specific conditional bits. VSI probably kept this.

For enhancements... Write your request on the back of a hundred-dollar
bill 💵 and mail it 💸 to OpenVMS World Headquarters 🌎 in Bolton Mass.
😜
--
Pure Personal Opinion | HoffmanLabs LLC
Rod Regier
2020-06-29 13:24:52 UTC
Permalink
I've used USB to SATA external enclosure products with OpenVMS/I64 systems.
You could put SATA SSds in there but you would *not* be taking advantage of the full performance the SSD could deliver.

The one advantage of an SSD over a magnetic disk in that content would be the faster "seek" performance in an SSD :-)

For most usage a fast USB "key" would do the trick for such usage and be less cumbersome to deploy.
Rod Regier
2020-08-11 14:20:58 UTC
Permalink
I found that the V/A 8.4 version of the MSA$UTIL didn't do a very good job of creating Units online.

I had to revert to using the ORCA firmware in the Pxxx controller
to create units on Alphas. (press F8 to...)

The V/I64 version of MSA$UTIL did a much better job dealing with Units.

HPE V/A 8.4 MSA$UTIL:

image name: "MSA$UTIL"
image file identification: "X-29"
image file build identification: "XCCV-0100000004"
link date/time: 27-MAY-2011 12:56:52.51
linker identification: "A13-04"

HPE V/I64 8.4 MSA$UTIL:

Image name: "MSA$UTIL"
Global Symbol Table name: "MSA$UTIL"
Image file identification: "X-35"
Image build identification: "0100000190"
Link identification: "Linker I02-37"
Link Date/Time: 4-MAR-2016 16:36:59.07
John H. Reinhardt
2020-08-11 14:40:48 UTC
Permalink
Post by Rod Regier
I found that the V/A 8.4 version of the MSA$UTIL didn't do a very good job of creating Units online.
I had to revert to using the ORCA firmware in the Pxxx controller
to create units on Alphas. (press F8 to...)
The V/I64 version of MSA$UTIL did a much better job dealing with Units.
image name: "MSA$UTIL"
image file identification: "X-29"
image file build identification: "XCCV-0100000004"
link date/time: 27-MAY-2011 12:56:52.51
linker identification: "A13-04"
Image name: "MSA$UTIL"
Global Symbol Table name: "MSA$UTIL"
Image file identification: "X-35"
Image build identification: "0100000190"
Link identification: "Linker I02-37"
Link Date/Time: 4-MAR-2016 16:36:59.07
I haven't been able to get the DS10 to run the P410 BIOS to get to ORCA. I'm using the serial console, not the graphics. Maybe that makes a difference. I don't have a graphics card in the system.
--
John H. Reinhardt
John H. Reinhardt
2020-08-11 15:08:49 UTC
Permalink
Post by Rod Regier
I found that the V/A 8.4 version of the MSA$UTIL didn't do a very good job of creating Units online.
I had to revert to using the ORCA firmware in the Pxxx controller
to create units on Alphas.  (press F8 to...)
The V/I64 version of MSA$UTIL did a much better job dealing with Units.
image name: "MSA$UTIL"
image file identification: "X-29"
image file build identification: "XCCV-0100000004"
link date/time: 27-MAY-2011 12:56:52.51
linker identification: "A13-04"
Image name:                                 "MSA$UTIL"
Global Symbol Table name:                   "MSA$UTIL"
Image file identification:                  "X-35"
Image build identification:                 "0100000190"
Link identification:                        "Linker I02-37"
Link Date/Time:                              4-MAR-2016 16:36:59.07
I haven't been able to get the DS10 to run the P410 BIOS to get to ORCA.  I'm using the serial console, not the graphics.  Maybe that makes a difference.  I don't have a graphics card in the system.
The DS10 sees the BIOS but has no way to get there.
Post by Rod Regier
show bios
Resetting I/O buses...
open fibre pga0.0.0.3004.0
open fibre pgb0.0.0.3005.0

pka0.7.0.15.0 - Adaptec AIC-7899
pkb0.7.0.115.0 - Adaptec AIC-7899
- 323A103C/3243103C <---- P410 no device name to run BIOS
--
John H. Reinhardt
Rod Regier
2020-08-11 15:17:22 UTC
Permalink
It's possible to use ORCA on "headless" system thru the serial console.

F8 becomes <ESC>8, etc.

This was the magic I used with the SA6402 on the DS15 FWIW:
(Tip of the hat to Stephen Hoffman)
set heap_expand 2mb
init
The SA6402 menuing and initialization is designed for video console operation and PC keyboard mappings. Switch DS15 to video console mode if not already in that mode.

There is no boot support for the SA6402 on a DS15 operated in SERIAL console mode!
set console graphic
run bios pya0
n.b. failure to provide the SA6402 with an expanded DS15 heap will result in the SA6402 failing to boot into its BIOS and omission of the logical disks delivered by the SA6402.

Logical disks created under the SA6402 are named DYAx offline, but tend to be named DKCx online.
set bootbios pya
INIT
set bootdef_dev DYA0
If you have the set bootdef_dev command complaining about DYA0 being an invalid device, try booting from DYA0 and then performing the SET command. Suspect there is a "state" issue with the SA6402 that interacts with DYA0 virtual device characteristics.
set bootdef_dev dya0.0.0.2004.2
SHO BOOT*
and copy the value associated with booted_dev
John H. Reinhardt
2020-08-11 16:12:53 UTC
Permalink
Post by Rod Regier
It's possible to use ORCA on "headless" system thru the serial console.
F8 becomes <ESC>8, etc.
(Tip of the hat to Stephen Hoffman)
set heap_expand 2mb
init
The SA6402 menuing and initialization is designed for video console operation and PC keyboard mappings. Switch DS15 to video console mode if not already in that mode.
There is no boot support for the SA6402 on a DS15 operated in SERIAL console mode!
set console graphic
run bios pya0
n.b. failure to provide the SA6402 with an expanded DS15 heap will result in the SA6402 failing to boot into its BIOS and omission of the logical disks delivered by the SA6402.
Logical disks created under the SA6402 are named DYAx offline, but tend to be named DKCx online.
set bootbios pya
INIT
set bootdef_dev DYA0
If you have the set bootdef_dev command complaining about DYA0 being an invalid device, try booting from DYA0 and then performing the SET command. Suspect there is a "state" issue with the SA6402 that interacts with DYA0 virtual device characteristics.
set bootdef_dev dya0.0.0.2004.2
SHO BOOT*
and copy the value associated with booted_dev
I've done all that with a Smart Array 5300. The 6400 won't work in a DS10 due to it being 3.3V only PCI.

But I'm not using an SA64xx or 53xx. I'm using a P410. No device shows up in the list because the SRM doesn't know about the P410.
Post by Rod Regier
show bios
Resetting I/O buses...
open fibre pga0.0.0.3004.0
open fibre pgb0.0.0.3005.0

pka0.7.0.15.0 - Adaptec AIC-7899
pkb0.7.0.115.0 - Adaptec AIC-7899
- 323A103C/3243103C <-- HP P410 note no device name
Post by Rod Regier
show config
AlphaServer DS10 617 MHz

SRM Console: V7.3-1
PALcode: OpenVMS PALcode V1.98-83, Tru64 UNIX PALcode V1.92-73

<snip>

PCI Hose 00
Bus 00 Slot 07: Acer Labs M1543C
Bridge to Bus 1, ISA
Bus 00 Slot 09: DE500-BA Network Controller
ewa0.0.0.9.0 00-10-64-30-25-AB
Bus 00 Slot 11: DE500-BA Network Controller
ewb0.0.0.11.0 00-10-64-30-25-A9
Bus 00 Slot 13: Acer Labs M1543C IDE
dqa.0.0.13.0
dqa0.0.0.13.0 Compaq CRD-8402B
Bus 00 Slot 15/0: Adaptec AIC-7899
pka0.7.0.15.0 SCSI Bus ID 7
dka100.1.0.15.0 FUJITSU MAP3367NC
dka200.2.0.15.0 FUJITSU MAP3367NC
Bus 00 Slot 15/1: Adaptec AIC-7899
pkb0.7.0.115.0 SCSI Bus ID 7
dkb0.0.0.115.0 CODESRC SCSI2SDV6
dkb100.1.0.115.0 CODESRC SCSI2SDV6
dkb200.2.0.115.0 CODESRC SCSI2SDV6
dkb300.3.0.115.0 CODESRC SCSI2SDV6
dkb400.4.0.115.0 CODESRC SCSI2SDV6
dkb500.5.0.115.0 CODESRC SCSI2SDV6
dkb600.6.0.115.0 CODESRC SCSI2SDV6
Bus 00 Slot 16: 811110B5
Bridge to Bus 2, PCI
Bus 00 Slot 17: Intel 21154-*E
Bridge to Bus 3, PCI
P410-> Bus 02 Slot 00: 323A103C/3243103C

Bus 03 Slot 04: FCA-2354
pga0.0.0.3004.0 WWN 1000-0000-c943-8
Bus 03 Slot 05: FCA-2354
pgb0.0.0.3005.0 WWN 1000-0000-c943-8
I think I will have to move the P410 and disks to my ZX6000 and hope that it works there to configure and then move the stuff back. THis means digging the ZX6000 out of storage and hoping it boots far enough to run ORCA
--
John H. Reinhardt
Rod Regier
2020-08-13 13:18:04 UTC
Permalink
P410 using ORCA should work on your I64.

If you want to talk to the P410 from OpenVMS/I64 you'll need to do something like this:

In order for the p410 to operate with OpenVMS/I64 including boot support the followup file must be included on the bootable system disk or any other boot media:

sys$system:SYS$USER_CONFIG.DAT:

device = "Smart Array P-series P400"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x3230103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT, BOOT
end_device
device = "Smart Array P-series PMC PCIe"
name = PK
driver = SYS$PKRDRIVER
adapter = PCI
id = 0x323A103C
boot_class = DK
boot_flags = HW_CTRL_LTR, UNIT_0
flags = CISS, PORT, BOOT
end_device


\\

P410i firmware update download:

https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX-a17b1b7dcc284cf5a21ecfa7c6#tab1

Transfer to FAT32 formatted USB key as "flat" directory structure.

Installing Package: Loading Firmware files onto EFI partition using USB.

Step: Download appropriate files to a USB key
Step: Place the USB key on the system to be updated
Proceed to "OS Independent - Updating Firmware from EFI shell", shown below.

OS Independent - Updating Firmware from EFI Shell
For general information about using EFI, refer to the "Operation and Maintenance Guide".

At the EFI shell prompt, type "map -r".
Then type "fs0:", replacing "fs0:" with each "fs" entry in the table until you find the volume containing the "fwupdate.nsh" file. We will assume this volume is labeled fs0 for the remainder of these instructions.
Type "fs0:", then press ENTER.
Locate the file "fwupdate.nsh" directory and type "fwupdate.nsh", then press ENTER.
NOTE:

In order to upgrade firmware make sure that all the required files are available in the same directory as “fwupdate.nsh”.

Reset the system: fs0:\EFI\hp\firmware>reset
John H. Reinhardt
2020-08-16 17:35:35 UTC
Permalink
Post by Rod Regier
P410 using ORCA should work on your I64.
Unfortunately, it doesn't. The P410 is in "HBA Mode" and won't start ORCA when the ZX6000 boots. So I found you have to set it to "RAID Mode" and then it works. The way to set it is to run saupdate. However I get this:

fs0:\> saupdate set_mode p410 raid

WARNING:Changing controller mode could result in DATA LOSS!!!
Key in 'YES' to continue or 'NO' to quit:yes

SAUPDATE supports changing of controller mode only for P410i and P411 controllers.The controller at 0:61:0:0 is not a P410i or a P411, hence its mode cannot be changed

fs0:\> saupdate list

********************************************************************************
Smart Array Offline Firmware Update Utility
Version 2.11.08.00

(C) Copyright 2009 Hewlett-Packard Development Company, L.P.
********************************************************************************

Searching and Connecting Smart Array Controllers...

********************************************************************************
Smart Array Offline Firmware Update Utility
Version 2.11.08.00

(C) Copyright 2009 Hewlett-Packard Development Company, L.P.
********************************************************************************

Seg Bus Dev Func Description Version Build
(WWN)

0 61 0 0 HP Smart Array P410 6.64 0
(0x0000000000000000)



fs0:\> saupdate get_mode 0:61:0:0

SAUPDATE supports querying of controller mode only for P410i and P411 controllers.The controller at 0:61:0:0 is not a P410i or a P411, hence its current mode cannot be obtained


So apparently saupdate can't handle a plain P410. I'm using the saupdate that comes with the firmware upgrade that you linked to for the 6.64 Smart Array.

Another post said that the drvcfg utility in EFI would do it so I tried that


T D
D Y C I
R P F A
V VERSION E G G #D #C DRIVER NAME IMAGE NAME
== ======== = = = == == =================================== ===================
11 00000010 B - - 7 13 PCI Bus Driver PciBus
1C 00000358 B X X 1 5 Smart Array SAS Driver v3.58 PciRom Seg=00000000

fs0:\> drvcfg -s 98 1C
drvcfg: Invalid handle 98

But that doesn't work either.

All of this is a moot point as now the ZX6000 will not start up. It runs for a about 45 seconds and powers back down with this message in the System Logs:

Waiting for new events...
01 Jan 1998 14:05:18
Alert Level 3: Warning
Keyword: Type-02 040601 263681
Fan speed lags
Logged by: Baseboard Management Controller;
Sensor: Fan - Fan 1R (Power)
Data1: Performance Lags
0x2034ABA29E020740 FFFF0106E8040300

01 Jan 1998 14:05:18
Alert Level 3: Warning
Keyword: Type-02 040601 263681
Fan speed lags
Logged by: Baseboard Management Controller;
Sensor: Fan - Fan 1L (Power)
Data1: Performance Lags
0x2034ABA29E020750 FFFF0106E9040300


So apparently I have an issue with the power supply fans. It has only one power supply so I'm not sure where the 1R and 1L come from. That nomenclature doesn't match the other fans in the chassis which are 1A, 1B, 2 and 3. I have two used power supplies on order but won't be here for a week.
--
John H. Reinhardt
John H. Reinhardt
2020-10-06 15:45:32 UTC
Permalink
Post by John H. Reinhardt
Post by Rod Regier
P410 using ORCA should work on your I64.
fs0:\> saupdate set_mode p410 raid
WARNING:Changing controller mode could result in DATA LOSS!!!
Key in 'YES' to continue or 'NO' to quit:yes
SAUPDATE supports changing of controller mode only for P410i and P411 controllers.The controller at 0:61:0:0 is not a P410i or a P411, hence its mode cannot be changed
fs0:\> saupdate list
********************************************************************************
                   Smart Array Offline Firmware Update Utility
                               Version 2.11.08.00
          (C) Copyright 2009 Hewlett-Packard Development Company, L.P.
********************************************************************************
 Searching and Connecting Smart Array Controllers...
********************************************************************************
                   Smart Array Offline Firmware Update Utility
                               Version 2.11.08.00
          (C) Copyright 2009 Hewlett-Packard Development Company, L.P.
********************************************************************************
         Seg  Bus  Dev  Func        Description          Version    Build
                                       (WWN)
          0   61    0    0       HP Smart Array P410      6.64         0
                                 (0x0000000000000000)
fs0:\> saupdate get_mode 0:61:0:0
SAUPDATE supports querying of controller mode only for P410i and P411 controllers.The controller at 0:61:0:0 is not a P410i or a P411, hence its current mode cannot be obtained
So apparently saupdate can't handle a plain P410.  I'm using the saupdate that comes with the firmware upgrade that you linked to for the 6.64 Smart Array.
Another post said that the drvcfg utility in EFI would do it so I tried that
            T   D
D           Y C I
R           P F A
V  VERSION  E G G #D #C DRIVER NAME                         IMAGE NAME
== ======== = = = == == =================================== ===================
11 00000010 B - -  7 13 PCI Bus Driver                       PciBus
1C 00000358 B X X  1  5 Smart Array SAS Driver v3.58         PciRom Seg=00000000
fs0:\> drvcfg -s 98 1C
drvcfg: Invalid handle 98
But that doesn't work either.
Waiting for new events...
01 Jan 1998 14:05:18
Alert Level 3: Warning
Keyword: Type-02 040601 263681
Fan speed lags
Logged by: Baseboard Management Controller;
           Sensor: Fan - Fan 1R (Power)
Data1: Performance Lags
0x2034ABA29E020740 FFFF0106E8040300
01 Jan 1998 14:05:18
Alert Level 3: Warning
Keyword: Type-02 040601 263681
Fan speed lags
Logged by: Baseboard Management Controller;
           Sensor: Fan - Fan 1L (Power)
Data1: Performance Lags
0x2034ABA29E020750 FFFF0106E9040300
So apparently I have an issue with the power supply fans.  It has only one power supply so I'm not sure where the 1R and 1L come from.  That nomenclature doesn't match the other fans in the chassis which are 1A, 1B, 2 and 3.  I have two used power supplies on order but won't be here for a week.
More on this saga. I got the new fans for the ZX6000. I'm wondering if I messed up the firmware because the tower version uses different fans than the rack version and when you switch between the two, you have to load different firmware. The new fans I got are for the rack version. Now the BMC is happy with the fans, but getting a CPU not starting error. So the ZX6000 is dead until I figure that problem out.
--
John H. Reinhardt
John H. Reinhardt
2020-10-06 16:52:09 UTC
Permalink
These days I'm looking for NOS 73G and above 15K 80pin 3.5 form factor SCSI.
Yup - unobtainium :-(
                     \\
I'm getting close to trialing a SATA SSD solution for RX2600 units w/OVMS boot support (new P410i PCIe controller and PCIX to PCIe adapter plus cabling).  The SATA drives will be deployed in an external enclosure.  I have a bill of materials drafted.  Parts are trickling in for the bench trial.
Alas, no Alpha boot support and even to use as a data disk requires 3.3V PCIx buss on DS15 and above.
https://groups.google.com/forum/#!topic/comp.os.vms/t7Mu4kEYTrw
Both the DS10 and the DS15 have the same PCI buss.  From the Technical Summary documents on both, it's a PCI V2.1 compliant bus supporting either %.0V or 3.3V PCI cards.
But all the PICe to PCI converter boards I see say they are PCI V2.3 compliant so I don't know if they would work with the DS10/15 PCI bus.
This one is common on Amazon in the US StarTech PCIPEX1 - <https://www.startech.com/Cards-Adapters/Slot-Extension/PCI-to-PCI-Express-Adapter-Card~PCI1PEX1>
Is there any chance a DS10 with V8.4 (HP, not VSI) could boot from a P400 or P800 series SATA HBA?
I think I'm at a final spot for this experiment.

With the ZX6000 logging a CPU failure, I took another tack. I searched for the cheapest used PC that had a PCI-e slot (and also PCI-X for another project) and found a Compaq 4000 Pro Xeon system. It recognized the P410 card, but unfortunately wouldn't run ORCA from the card BIOS. Googling showed that you need an HP Proliant system BIOS to do it. Frack. Should have researched better. So I found the cheapest HP Proliant server - an ML110. I found a ML110 G4 that had a combination of two PCI-X and two PCI-E slots so I bought that. Turns out it's too old. The BIOS also recognizes the P410 but won't run ORCA. So I Google some more and out the first ML110 to support the P410 Smart Array is a ML110 G7. Another Ebay purchase and I've got it and it does work! But it's only got PCI-E slots and it's 4" deeper than the G4 and about 30lbs heavier. So I go searching again and find the online version of ORCA works with Linux or Windows. I have an eval copy of Windows server 2019 so I put it on the ML110 G4. But it's horribly slow. The G4 is a Pentium D dual core 3Ghz CPU. Only 1GB of memory so I find a 4GB set and install it. The HP SSA Raid Utility works (with a god-awful tablet style GUI - on a server O/S really???) But it works. I can configure the P410 and do everything I need.

To facilitate configuring, I bought an identical ICY Dock SATA chassis for the ML110. So when configuring I pop the drives out of the chassis in the DS10 and put them in the chassis in the ML110, transfer the P410 Smart Array card to the ML110 and plug in the two SATA cables. Now it's set up exactly as in the DS10 and the HP SSA Raid configuration tool can create a configuration that works in both.

This setup probably isn't real cost effective but it was an interesting experiment and still cheaper than buying ACARD SCSI-to-SATA adapters at $280 for each drive. The HP P410's (with a 1GB flash-based write cache) are between $30 and $50 on Ebay so relatively cheap. The ICY Dock chassis will accommodate either SATA hard drives or SATA SSDs

HP P420 Smart Array cards are also pretty cheap and allow better RAID 5 and RAID 6 performance in addition to being 6GB SATA as compared to the 3GB SATA of the P410. This is probably only an advantage when doing internal card transfers as the PCI interface would be swamped by even a 3GB transfer speed. I only plan to mirror the drives in a RAID 1 so I'm not looking too hard at the P420 card. I have to decide if I will buy a 3rd SSD to make a hot spare if/when one of the main SSD drives fails. If not, then if a failure happens I would have to move everything to the ML110 and rebuild there with the replacement drive. But the SSD may not fail for a long time as I doubt I'll be pushing the TBW rating on it.

The Samsung EVO 860 SSD have been running good for about a month now. My biggest concern with them is handling of garbage collection and trim on unused space. I don't think the P410 firmware has much for handling SSD drives in it. The volumes I've created in OpenVMS have highwater marking turned on as a result of seeing some things mentioned here on C.O.V. about that being the trigger for OpenVMS to do at least some space management with trim.

As it stands now this is the configuration:
DS10 AlphaServer
HP P410 Smart Array SATA HBA
PLX8112 based PCI to PCI-E bridge adapter <https://www.ebay.com/itm/PCI-TO-PCIe-Bridge-Card-CHIPSET-PLX8112/233387975013>
Also available on Amazon at a higher price but faster delivery. Do NOT get the StarTech. The PCI-E slot on it won't accomodate the P410
<https://www.amazon.com/dp/B07ZX9B36Y>
Mini SAS to SATA Cable (SFF-8087 to SATA Forward Breakout) 1.6 Feet (Two needed for all 6 bays. One if you only want 4 drives)
<https://www.amazon.com/gp/product/B018YHS8BS>
ICY DOCK ToughArmor MB608SP-B Rugged Full Metal 6 Bay 2.5" SATA HDD & SSD Removable Drive Enclosure
<https://www.amazon.com/gp/product/B07DM2NQNW>
Samsung SSD 860 EVO 250GB SSD
<https://www.amazon.com/gp/product/B07864WMK8>

Pros:
Newer, faster, more reliable SATA or SATA SSD drives
RAID/Mirror on Smart Array works as fast as disks will allow (3GB SATA speeds)
Larger capacity drives available

Cons:
PCI/PCI-e bottleneck limits speed but still as fast as 1990's SCSI allows.
MSA$UTIL can't configure so have to move drives and controller to another system to configure


Anyone want a cheap Compaq 40
Rod Regier
2020-10-06 17:27:34 UTC
Permalink
JHR, I can't find any major faults with the approach you have taken.

How are you mounting/powering the Icy Dock multi-drive Removable Drive Enclosure Backplane Cage for 5.25" Bay?
John H. Reinhardt
2020-10-06 18:17:34 UTC
Permalink
Post by Rod Regier
JHR, I can't find any major faults with the approach you have taken.
How are you mounting/powering the Icy Dock multi-drive Removable Drive Enclosure Backplane Cage for 5.25" Bay?
Rod,

It fits in one of the two 5.25" spots in the DS10. It has two SATA power sockets on the back so I just used a molex to SATA splitter.

<https://www.amazon.com/gp/product/B00VJ9V8NY>
--
John H. Reinhardt
Rod Regier
2020-10-06 18:34:55 UTC
Permalink
Lucky fellow.

No such apparent options on the RX2600. (Optical drive bay is a dedicated slimline). No Molex power to "steal". Only way I can figure out to readily provision 3+ 2.5 SATA SSD drives on an RX2600 is to use an *external* standalone SATA 2.5 multi-bay enclosure with it's own power supply. That runs up the price fast :-(

On the RX2600 the 3x 3.5 80P SCSI drive cage sheet metal is riveted in place.
John H. Reinhardt
2020-10-06 19:01:04 UTC
Permalink
Post by Rod Regier
Lucky fellow.
No such apparent options on the RX2600. (Optical drive bay is a dedicated slimline). No Molex power to "steal". Only way I can figure out to readily provision 3+ 2.5 SATA SSD drives on an RX2600 is to use an *external* standalone SATA 2.5 multi-bay enclosure with it's own power supply. That runs up the price fast :-(
On the RX2600 the 3x 3.5 80P SCSI drive cage sheet metal is riveted in place.
Yep. I noticed that when working with my ZX6000 which is basically an RX2600 in a tower case. When I was attempting to do the P410 configuration on it I had to supply power with an external brick that happened to have a molex connector. Definitely not handy. It's one thing that convinced me to get the Proliant ML110 to do the configuration as it was so much more convenient.

The only way I can see to do SATA on the RX2600/2620/ZX6000 is a Smart Array card with an external port such as a P600 and an external disk enclosure as you mentioned.
--
John H. Reinhardt
Rod Regier
2020-10-13 13:31:19 UTC
Permalink
Post by John H. Reinhardt
The only way I can see to do SATA on the RX2600/2620/ZX6000 is a Smart Array card with an external port such as a P600 and an external disk enclosure as you mentioned.
I had earlier published a bill of materials on how a P410i card can talk to an external eSATA standalone enclosure. (uses the P410i internal connector) w/o SKUs it would look like:
* PCIX to PCIe adapter
*P410
*P410 1G cache card and FBWC
*internal/internal 4x cable
*Int to ext 4x chassis adapter (hard to source)
*External/external 4x single cable
*Standalone eSATA enclosure (moderately hard to source)
*1-4 SATA drives

I have found suppliers and SKUs for all of the above.
John H. Reinhardt
2020-10-13 17:03:25 UTC
Permalink
Post by Rod Regier
Post by John H. Reinhardt
The only way I can see to do SATA on the RX2600/2620/ZX6000 is a Smart Array card with an external port such as a P600 and an external disk enclosure as you mentioned.
* PCIX to PCIe adapter
*P410
*P410 1G cache card and FBWC
*internal/internal 4x cable
*Int to ext 4x chassis adapter (hard to source)
*External/external 4x single cable
*Standalone eSATA enclosure (moderately hard to source)
*1-4 SATA drives
I have found suppliers and SKUs for all of the above.
True. I guess my point was you need an external box. It's matter of choice and availability on how you get there.
--
John H. Reinhardt
Loading...