HP OpenVMS Installation problems on Itanium2-Server (rx2620)

[Alexander Schnapper]

Hello,

I have got a problem installing HP OpenVMS 8.2 onto a hp rx2620 server.
At the beginning of the installation I need to provide the setup with the target disk device name.
But when I enter '?' for more choices I see the other disks as 'offline' and I can't install it.

The disks are empty, no system on it.

Has anyone an idea?

Thank you in advance,
Best Regards,
Alexander Schnapper

[Volker Halle]

Alexander,

what kind of disks are these and how are they connected ?

Do you see the disks as blk devices using the MAP command at the EFI shell prompt ?

Volker.

[Alexander Schnapper]

Hello Volker,

thanks for your answer.
They are two 36GB 15k HotPlug SCSI Disks.

> Do you see the disks as blk devices using
> the MAP command at the EFI shell prompt ?

Yes. And the disks are recognised at startup.

I can install Linux, HP-UX or even Windows on those disks (I tried) and that's no problem, so the disks work.

Any idea?

Thanks,
Alex

[David B Sneddon]

Alexander,

Just a thought... do you have the latest firmware?

Dave

[Volker Halle]

Alexander,

the devices should be DKA0 and DKA100, right ?

Did you check the Firmware Version ?

Shell> info fw

should give

Firmware Revision: 2.13 [4412]

When OpenVMS is configuring the devices, there should be PKA0 and PKB0 messages showing the operating mode and controller firmware versions, e.g.:

%PKA0, Copyright (c) 2001 LSI Logic, PKM V1.1.01
%PKA0, SCSI Chip is LSI53C1030, Operating mode is LVD Ultra320 SCSI
%PKA0, LSI53C1030 firmware version is 1.3.35.0
...

You could enter DCL (menue #7) and look around for something unusual:

$ MC SYSMAN IO AUTO/LOG
$ ANAL/SYS
SDA> CLUE CONFIG
SDA> CLUE SCSI/SUMM
SDA> EXIT
$ scsi_info:==$sys$etc:SCSI_INFO
$ scsi_info dka0
$ show error
$ INIT DKA0: test
...

Volker.

[Alexander Schnapper]

Hi Dave,

You wrote:
> Just a thought... do you have the latest
> firmware?

Yes, if you mean the MP I've got MP Version 3.13 and Firmware 3.10.

Thanks :-)
Alex

[Ian Miller.]

can you describe your configuration in more detail? To which scsi bus are the disks attached?

[Alexander Schnapper]

Hi Volker,

You wrote:
> the devices should be DKA0 and DKA100, right ?

> Did you check the Firmware Version ?

Yes, my Firmware Version is 3.13


> When OpenVMS is configuring the devices, there should be PKA0 and PKB0 messages showing the operating mode and controller firmware versions, e.g.:

> %PKA0, Copyright (c) 2001 LSI Logic, PKM V1.1.01
> %PKA0, SCSI Chip is LSI53C1030, Operating mode is LVD Ultra320 SCSI
> %PKA0, LSI53C1030 firmware version is 1.3.35.0
> ...

Yes, thats true. The output even gives me

%PKA0 .... firmware 1.3.35.65
%PKB0 .... firmwrae 1.3.35.65

> You could enter DCL (menue #7) and look around for something unusual:

> $ MC SYSMAN IO AUTO/LOG
> $ ANAL/SYS
> SDA> CLUE CONFIG
> SDA> CLUE SCSI/SUMM

Output is for PKB0 nothing, but for PKA0 it says:

'DKA0 .....
............
DKA100'

> SDA> EXIT
> $ scsi_info:==$sys$etc:SCSI_INFO
> $ scsi_info dka0

Output from this command was this:

'
SCSI_INFO V3
Issuing INQUIRY QIOW
QIO status of 132 causing program exit
%SYSTEM-F-DEVOFFLINE, device is not in configuration or not available
'

> $ show error

Output is:

'
Device Error Count
DQA1: 2
DKA0: 11
DKA100: 11
'

> $ INIT DKA0: test

Output is:

'
%INIT-F-DEVOFFLINE, device is not in configuration or not available
'

Thanks so far again for your help.

[Volker Halle]

Alexander,

I've recently set up an OpenVMS V8.2 rx2620 system with this same system FW rev 3.10 and config (just one device DKA0 - PKA0 also showing FW 1.3.35.65), so it should not be a general problem.

Could this be some setup problem with the SCSI adapter ? Could you just remove one disk and/or plug it into the 3rd slot ?

The error log information might be helpful, but how to save it with no writeable media ?!

Volker.

[Volker Halle]

Alexander,

you could try the unsupported DKLOG SDA extension:

$$$ SET TERM/WID=132
$$$ ANAL/SYS
SDA> DKLOG START DKA0
SDA> SPAWN MC SYSMAN IO SCSI_PATH_VERIFY
SDA> SPAWN MC SYSMAN IO AUTO/LOG
SDA> DKLOG STOP DKA0
SDA> DKLOG SHOW DKA0
SDA> EXIT

You would be looking for anything likely to indicate an error...

Volker.

[Alexander Schnapper]

Hello Volker,

you wrote:
> you could try the unsupported DKLOG SDA extension:

> $$$ SET TERM/WID=132
> $$$ ANAL/SYS
> SDA> DKLOG START DKA0

Output was:
'
Starting logging on specified devices...
... completed after 1 UCB with 0 no-buffer errors
'

SDA> SPAWN MC SYSMAN IO SCSI_PATH_VERIFY
SDA> SPAWN MC SYSMAN IO AUTO/LOG
SDA> DKLOG STOP DKA0

Output was:
'
Stopping logging on specified devices...
... completed after 1 UCB with 0 no-buffer errors
'
SDA> DKLOG SHOW DKA0

A Table was shown after this command with Log Points (SENSE_DATA and CMD_BEGUN, CMD_ENDED) and a SCSI Statur (CHK COND).

SDA> EXIT

> You would be looking for anything likely to indicate an error...

Hm, I couldn't indicate an error...



The disks are connected to the SCSI-Controller of the system.


I try to install it again, lets hope it works....

Thanks, Best Regards,
Alexander

[Volker Halle]

Alexander,

SCSI CHK COND and SENSE_DATA seem to indicate SCSI errors/problems ! Could you post the information in a .TXT attachment ?
The errorlog entries will be stored in memory, but I can't think of an easy way to decode them. SDA> CLUE ERRLOG will show them, but can't write CLUE$ERRLOG.SYS.

Don't bother to try another installation, if you can't access the disks DKA0/DKA100 using DCL (e.g. INIT) when booted from the CD.

Did you try to move the disks into another slot ?

Volker.

[Volker Halle]

Alexander,

it is possible to look at the errlog buffers in memory using SDA, but decoding them may be quite hard...

$$$ ANAL/SYS
SDA> EXA @exe$al_erlbufadr;10
00000000 88035000 88034080 00000000 ...
XXXXXXXX ^endofbf ... ^nextbfr ... ^flags+cnt

SDA> EXA @exe$al_erlbufadr+80;100

will show the first errlog buffer. You'll see SOME ASCII text.

SDA> exa @exe$al_erlbufadr;@exe$GB_ERLBUFPAGELETS*200

will dump the contents of the first errlog allocation buffer - may be quite lengthy - but you should see DKA0 and DKA100 errlog entries (the device names are shown in ASCII).

Once such an errlog entry is located, it needs to be decoded to find the ASC/ASCQ SCSI sense codes and their meaning...

Volker.

[Alexander Schnapper]

Hello all,

thanks for your support.

It worked, but guess what the problem was?!

The SCSI-Cable wasn't put in properly. So perhaps while the installation processes of other OSes could mount the drives, OpenVMS couldn't (time-out connection).

Thanks for your support, have a nice week.

Alex

[Alexander Schnapper]

see upper comment