Devices

[Asim_5]

Hi

we have GS80 Alpha servers which is connected with MA8000 (HSG80 controller) storage. when we give command on SRM prompt of Alpha server GS80

p00>>show dev d

we are not able to see all the disk. but when the system boots we can see all the disks.

to see the disks on SRM prompt do we have to set any perameters on HSG80 controllers ?

[Uwe Zessin]

Asim,
the SRM console does not generate device names for fibre channel devices itself. You need to use the 'WWIDMGR' utility to lookup the devices.

I don't know if you need to set the console of the GS80 into a special diagnostic mode with 'set mode diag' before you can run WWIDMGR - please try it out.

The command '>>> wwidmgr -show wwid' will list all disk and tape LUNs that the server can see. Append a '-full' and you see even more details.

There should be a .PDF file on the AlphaServer firmware CD-ROM that descriptes the WWIDMGR commands.

[Mike Naime]

You will not see the disk devices on the SRM prompt {SLV1> for slave#1, I forget the master prompt}

you can see the devices at the "Console" prompt when you exit the SRM. This is your "P00>>" prompt on your GS80.

Question? Why are you concerned about seeing anything beyond the boot LUN at the console level? After the system is up and running, you can add/remove/modify disks on the fly.

If you see all the drives when you boot, there is nothing for you to do at the HSG level.

[Uwe Zessin]

Mike,
I thought the SLV> prompt is from the SCM (System Control Manager) level, that is from the nodes on the CSB (Console Serial Bus). The SRM is from the primary CPU. I have done a quick check in an old training manual for GS series and it looks like the prompt is 'SCM_E0>', where E0 is the CSB address.

I try to always verify at the console level already to make sure that the server has all access paths to a LUN since I once had to diagnose a system why it could no longer boot (zoning error in one fabric, missing boot path set through the other). There was some more history involved, but I guess you understand the reason.

[Mike Naime]

SCM/SRM/MBM/QBB/SBB... Etc! It all runs together after a while. With multiple names for the same function depending on the hardware type... I just call a console prompt with the >>> a console prompt. Anything else I think of as the Alphabet soup prompt. :-)

As I was explaining to our local/regional field service guys while installing a 32P 1280 and 2 full EVA5000's today, If I only worry about setting my boot LUN, I can verify everything else from the OS level. (IO AUTO /LO) Plus, because the EVA is an ALL or nothing proposition for the WWID's defined for the HOST. It follows that if my boot LUN is on the same EVA storage controller as the other LUNS, I know that I have my 4 paths to the other drives if all 4 paths show on the boot LUN. This is not selective to each individual LUN on the EVA.

With the HSG's I had to verify my connection names prior to Storage presentation and setting the boot LUN in WWIDMGR.

I know about tracing the Boot problem. The usual source of the problem is the HSG80 storage controller. Not the zoning, or the WWID settings on the Alphaserver. If re-booting the storage controller is not an option because of the other LIVE systems onm the storage controller, You can cure the symptoms by removing and re-adding the zones from the config. This causes the storage controller to re-set.

Mike Naime

[Uwe Zessin]

Ah, I see. You have another way to create missing connections on the HSG ;-)

I usually disable/enable the switch port of the affected FCA. I try to avoid zoning changes 'for fun' as I once crashed a Tru64 Unix system - the message was something like 'RSCN too small', so I suspect a bug in the OS.

[Jan van den Ende]

Well Uwe,

if by that you mean in the SAN OS, then I concur.
We have also "lived in interesting times" due to some HSG issues.!

Jan

[Uwe Zessin]

Hello Jan,

I was talking about the Brocade fibre channel switches where the firmware, as far as I know, is called 'Fabric OS'. The only 'SANOS' I am aware of is used on the CASA:
http://www.hp.com/products1/storage/products/virtualization_appliances/network/sv3000/index.html
I have only played with that box a few minutes during a technical pre-sales training about 2 years ago.

Anyway, I always like to learn new things and I would love to listen to you (honestly).

[Jan van den Ende]

Uwe,

guess I can't be of little help there.
In this organisation | Storage | VMS | *NIX | Citrix | Netware | "coexist", but are VERY separated, having their own responsabilities, bound to cooperate, but certainly NOT to interfere with each other's work.As far as I know "something" runs the HSG80's. and I loosely called that an OS, guess it's what you call firmware.
I DO remember that they needed to install patches, and that that SHOULD go unnoticed...
It took out the Tru64 stuff, and reduced us to single-member shadow sets and MANY ours of shadowcopying.

Jan

[Uwe Zessin]

The name for the operating system on the HSG storage controllers is called 'ACS' - Array Controller Software.

On most older HSZ-based controllers it is 'HSOF' - Hierarchical Storage Operating Firmware.

The firmware on the HSZ-10 is called ACS, too.

In case that isn't enough: the software for the old HSC-50 (delivered on TU58 tapes, ugh) is named CRONIC (Colorado Realtime Operating Nucleus [for] Intelligent Controllers).


Welcome to the club of HSG victims. On the other hand I am really happy that the controller is rather stable as long as you don't touch it or annoy it with unusual requests.