rx2600 won't boot VMS V8.3 from MSA1000

[Galen Tackett]

I've had VMS V8.2-1 running on an rx2600 since shortly after V8.2-1 shipped. The system disk, $1$DGA1:, is a raid 5 unit on an MSA1000.

I attempted to upgrade this system to VMS V8.3. The installation completed normally and the system shut down as expected.

I started it back up and paused at the Boot Options menu. As expected, the installation procedure had re-created the boot options for $1$DGA1:

When I went to boot the upgraded disk I got this:

Loading.: VMS V8.3 ON MSA1000 FGB0.5008-05F3-000D-DDD1
Starting: VMS V8.3 ON MSA1000 FGB0.5008-05F3-000D-DDD1

So far, so good. Now it just sits for a minute or so, and then:

2 0 0x000285 0x0...00 SAL_CHECK Initialized Store
7 0 0x000098 0x0...0B Machine Check Initiated
2 0 0x0002a1 0x28000000fff21130 Processor PSP
2 0 0x000115 0x0...0 MCA: Uncorrected Machine Check
2 0 0x0002b0 0x2006112100092408 MC: MC Timestamp
2 0 0x00028a 0x0...0 MC: Logging Processor State
2 0 0x000282 0x0...0 Processor Cache Check
2 0 0x000281 0x1080000003000340 Processor Bus Check
2 0 0x00028e 0x0000004100000000 Processor Mod Err Info Target ID
2 0 0x000284 0x000000003f930d40 Processor IIP
2 0 0x000287 0x0...0 Processor IPSR
2 0 0x000289 0x0...0 MC: Logging Memory CEC
2 0 0x000288 0x0...0 MC: Logging I/O CEC
2 0 0x0002a0 0x0...0 MC Post Process Platform Log


After which I had to use the MP to reset the system.

I tried deleting the boot options and entering them both manually and via BOOT_OPTIONS.COM from the install DVD. Same error.

Booting from the installation DVD, I used BACKUP to copy the disk to one of the rx2600 internal SCSI drives. Then I set up an entry in the boot options menu to boot from the SCSI drive.

It boots just fine from the SCSI drive.

Again using the install DVD I then copied the SCSI drive back to $1$DGA1 and (again) created boot options menu entries for it.

Just for "fun" I tried a fresh install (as opposed to an upgrade) of V8.3 on a brand new unit ($1$DGA3) on the MSA1000. I get the same error. (Oddly, the installation created FOUR boot options men entries, the first two of which merely tell me "not found" if I try to use them.)

I copied my backup of VMS V8.2-1 back to the MSA1000 and was able to boot it okay.

So I would hazard a guess that the MSA1000 and VMS V8.3 aren't playing nice with each other.

This system isn't under any kind of support agreement and due to bureaucratic issues within our government customer it isn't likely to be under one any time soon.

Here are some particulars on the firmware:

rx2600 SYSREV reports:

MP FW : E.03.15
BMC FW : 01.52
EFI FW : 01.22
SYSTEM FW : 02.31

msa1000 software 4.32 build 300 hardware 7

[Galen Tackett]

I've installed all the currently available ECO kits for V8.3, by the way...

[Volker Halle]

Galen,

what's the EFI driver and firmware version for the HBA (A6826A) ?

I recently installed a 3 node rx4640 cluster booting via an EVA3000 using Driver 1.49 (firmware 3.03.154) on the HBAs.

I also booted an rx2620 about a year ago with HBA (driver 1.42, firmware 3.03.150) from an EVA4000 with the following FW:

MP FW E.03.13
BMC_FW 03.47
EFI FW 03.10
System FW 03.10

According to the OpenVMS V8.3 Release notes, your BMC FW seems to be one version too low:

rx2600 system: 2.31 BMC: 1.53 MP: E.03.15

I was just checking the revs. This should not imply, that this has anything to do with your machine check.

Volker.

[Richard Whalen]

I have succesfully installed 8.3 on a RX2600 with:
MP FW : E.03.15
BMC FW : 01.53
EFI FW : 01.22
System FW : 02.31
My system does not have a fibre channel adapter, so checking the firmware revision on that would also be in order.

[Volker Halle]

Galen,

if you have both HBAs connected to the MSA1000, it's o.k. for the installation procedure to create 4 boot entries. As the MSA1000 can't do active/active, you can only use 2 of those entries (both on the same HBA FGB).

I've had the same effect with the EVA3000, which was using old firmware and couldn't do active/active as well.

Volker.

[Galen Tackett]

I did notice the BMC FW being out of rev. but the release notes for 01.53 didn't mention anything remotely similar to this. As Volker says, of course, the BMC rev. may have nothing to do with this.

I've been trying to determine the firmware rev. on the A6826A Fibre Channel adapter but haven't found how to do this yet. If someone can tell me how, I'll check it tomorrow. (I only made a brief search today.)

[Volker Halle · Accepted solution]

Galen,

to determine (and eventually upgrade) the firmware of the HBA, you need the IPF Offline Diagnostics CD. Then run:

fs0:\EFI\HP\TOOLS\IO_CARDS\FC2> efiutil info

Fibre Channel Card Efi Utility 2.17 (03/23/2006)

2 Fibre Channel Adapters found:

Adapter Path WWN Driver (Firmware)
A0 Acpi(000222F0,600)/Pci(1|0) 50060B0000399BE0 1.49 (3.03.154)
A1 Acpi(000222F0,600)/Pci(1|1) 50060B0000399BE2 1.49 (3.03.154)

You will also see the following messages after initializing the rx2600, while it loads it's drivers:

...
HP 2Port 2Gb Fibre Channel Adapter (driver 1.49, firmware 3.03.154)
HP 2Port 2Gb Fibre Channel Adapter (driver 1.49, firmware 3.03.154)
...

Volker.

[R. Verkerk]

Hello,

we are bootting an Rx2620 from the MSA1000 with OpenVMS 8.3.

Our MSA1000 has firmware V4.48

with clue config:
System Management BIOS Information:
BIOS Version 04.10 Manufacturer hp
Release Date 08/01/2006 hp Product Name server rx2620

greetings,

Robert Verkerk

[Galen Tackett]

I just discovered this today in the VMS V8.3 Upgrade and Installation Manual:
"OpenVMS Version 8.3 for Integrity servers requires that the HP A6826A 2 GB Fibre Channel host-based adapter and its variants have the following minimum version: EFI driver: 1.47 RISC firmware: 3.03.154; HP AB378A and AB379A 4 GB Fibre Channel host-based adapter have the following minimum version: EFI driver: 1.05 RISC firmware: 4.00.70."

(see http://h71000.www7.hp.com/doc/83final/6677/6677pro_hardware2.html)

Once I get the Offline Diagnostics and Utilities CD burned and try it out, I'll report back.

[Galen Tackett]

I updated the A6826A firmware and am now able to boot from $1$DGAxxx devices.

[Ian Miller.]

A case of if all else fails read the manual:-)

[Galen Tackett]

> A case of if all else fails read the manual:-)

You know, Ian, I thought I did that at some point. It's possible I missed the important bit; also that age is taking its toll on my memory.

Uh... What was the subject here?

:-)

Thanks to all for your help, and to Ian for a pleasant smile.