KGPSA-BC card with VMS ver 7.2-1

[Vikram Sachdeva]

Hello

We are trying to install KGPSA-BC(32 bit fiber channel card) in DS20E to connect HSG80 controller via SAN switch 2/8 EL .

We are able to see the disks at SRM console level but not at OS level(VMS ver 7.2-1).I have searched through ITRC and HP website and its written there that KGPSA-BC card is supported with Patch/TIMA kit on VMS ver 7.2-1

Here are some facts
1.All the relevant patches (v3.0 Update patch and Fiber_SCSI patch )installed.
2.at HSG80 level SCSI_version is set to SCSI-3
and Command Console Lun (CCL) is also enabled
3.All the units created is assigned to unique IDENTIFIER .
4.in show connections OS_TYPE is set to VMS
5.We are able to see all the disks if we boot through VMS Ver 7.3-1 CD.
6.We can't upgrade to VMS ver 7.3-1 as application is not supported on this version
8.Hardware is Alpha Server DS20E
9.SRM Firmware : 6.2-1
10.Storage is MA8K with ACS 8.7P
11.KGPSA-BC card is set to FABRIC mode and we can see the disk at SRM level if we set UDID .

Please suggest what can be done so that we can see HSG80 disk at OS level.


Regards
Vikram Sachdeva

[Veli Körkkö]

w.r.t

>1.All the relevant patches (v3.0 Update patch and Fiber_SCSI patch )installed.

Please install all remaning patches also.

>2.at HSG80 level SCSI_version is set to SCSI-3 and Command Console Lun (CCL) is also enabled

Is the HSG80 set to MULTIBUS FAILOVER too?

...

>5.We are able to see all the disks if we boot through VMS Ver 7.3-1 CD.
>6.We can't upgrade to VMS ver 7.3-1 as application is not supported on this version

Could you consider VMS V7.2-2?

...

>9.SRM Firmware : 6.2-1

Upgrade preferably to latest SRM console, V6.6 is latest available AFAIK.

>10.Storage is MA8K with ACS 8.7P

Do you have the ACS V8.7 patches installed?
Do you really have V8.7P and not V8.7P-1?


>11.KGPSA-BC card is set to FABRIC mode and we can see the disk at SRM level if we set UDID

What do you exactly mean here? One would assume that

>>> wwidmgr -show wwid

shows all disks and here you mean maybe that if one does

>>> wwidmgr -quickset -udid xxx
>>> init

and you can then see the said lun using simply

>>> SHOW DEVICE

Is that so?

Are those switches the newer 2GB ones? If so,
maybe forcing the switch port down to 1GB
fixed might help, since KGPSA-BC is old 1GB card and there just might be link setup issue
between the switch and HBA when VMS tries to boot. Also consider upgrading the HBA firmware to latest level.

_veli

[Art Wiens]

Veli's comment on point 6 is the key, I believe. If I recall correctly, v7.2-2 with a "tima" kit or patch was required for fibre channel. I was able to restore a v7.2-1 image onto an HSG80 disk w/ a DS10 and a v7.3 cd, but could not boot it even though it was a current patched v7.2-1 disk image. I don't think v7.2-2 should give you any issues if your at v7.2-1.

HTH,
Art

[Benjamin Levy]

We have been running KGPSA-BC with VMS V7.2-1 and UPDATE-100 since summer-2000 without significant problems.

We run ACS V8.5.

One difference of note - we use the "F" series of ACS firmware. Not sure what the difference is between that and "P"

[Mike Naime]

When I'm setting up my Alpha's to boot from the SAN.

INIT

WWIDMGR -QUICKSET -UDID {BOOT LUN}
INIT

SET FFAUTO ON
SET BOOTDEF_DEV DG*{LUN}.*.*.*.*
SET DUMP_DEV DG*{LUN}.*.*.*.*

SHO BOOT*
you should show the 4 paths to your boot LUN.

We ran our DS20's at 7.2-1h1 for a couple of years on HSG80's/SAN

Have you tried updating the firmware in the HBA's? I'm assuming that the -BC is the LP7000 adapter. I know that the -CA/CX is the LP8000.

[Tim Nelson]

Here is the solution ( I am an associate of Kirk)
Seems the storage works FC switches and kgpsa-bc do not auto negotiate very well..
Sometimes they negotiate 1GB G sometimes 2GB F sometimes 1GB F etc.. Once the driver loads from VMS it renegotiates correctly, albeit too late in the boot sequence though. Hence the reason io autoconfig always worked after the fact but not during boot because the scan was timing out due to the bad negotiation.
We hard set the switch port to 1GB and now we can successfully see all storage on the SAN from the SRM level. The connection to the other SAN that did not seem to have a problem already had the ports hard set to 1GB F.

Thanks to all for your responses.

[Tim Nelson]

Please ignore my previous response. I responded in the wrong window for a different probelm.