Subject: RE: Re: KFQSA Resolution - at last
To: Douglas Meade <inforum@umd5.umd.edu>
From: Chuck McManis <cmcmanis@freegate.com>
List: port-vax
Date: 02/16/2000 10:27:34
Hello Doug and Port-Vax,

At 11:15 AM 2/16/00 -0500, Douglas Meade wrote:
>Well, again, nothing quite as optimistic as this old subject
>line suggests, but I think I'm almost there.  Your posting
>with the above title has proved a godsend to me, perhaps
>we should team up and make a mini-FAQ of it.

This is probably a good idea, we might even be able to convince Dave to put 
it up on the Hardware FAQ.

>I've been trying to get an "Industrial Vax 630" converted to
>a MicroVax III (3500?).  The Industrial Vax was a MicroVax II
>in a BA213 chassis, with 3 RD54 drives and a TK50.

A KA650 in a BA213 would be a MicroVAX 3500, a KA655 in a BA213 is a 
MicroVAX 3800.

>...  The first was
>due to the 10-pin connector that goes to the top of the control
>panel.  (This is the cable that runs to a connector behind the
>right hand side power supply.)

This attaches the control panel to the backplane I believe (and makes it 
possible to assert RESET and HALT on the Qbus.) If it was polarized I'm 
suprised that the connector wasn't polarized as well. DEC is very good 
about those things. Perhaps there was change from the MVII.

>ra0 at mscpbus 1   drive 8:  RF72
>
>(I'm curious why it's drive 8)


DSSI drives are "Intelligent Storage Elements" (ISEs) that actually have a 
controller built right into the drive. This makes the DSSI bus much more 
like the IBM DASD Channels than say ESDI or even SCSI. When you cluster two 
VAXen with the DSSI bus they both get access to the drives and the drive 
arbitrates between them. Anyway, one of the "features" of the drive is that 
it can either take its Unit number from the plug (which sets some pins on 
the 10 pin connector) or it can use the unit number it has stored in NVRAM 
on the controller. What it does is controlled by the FORCEID property in 
the MSCP setup screen.
When the controller is in program mode, the drive 0 plug will specify the 
drive you wish to talk to, then you can type SET HOST/MAINT/MSCP 0 talk to 
it and set its parameters. It will also tell you some interesting things 
about the drive such as how many times it has been power cycled, how many 
hours it has been running etc.

>I can do "disklabel /dev/ra0" and get a listing to the screen.
>I can also do "disklabel -W /dev/ra0" and "disklabel -B /dev/ra0"
>
>But when I actually edit a file of the disk structure (called "ra0")
>and try to do "disklabel -R -r /dev/ra0 ra0" I get the messages
>"write error", and "wtfs undefined".

Do you have this backwards? 'disklabel -W' writes so that last paragraph is 
broken right? Assuming it is, the message you got wasn't from the disk it 
was from your disktab entry. The complaint is that the parameter 'wtfs' 
isn't defined as a valid disktab parameter. If disklabel -R works, try 
disklabel -i and see if you can write the label out that way. It also makes 
computation easier.

>I've got two other disks I can try, but I didn't get around to it
>last night.

You are very nearly their I suspect.

>My question is, is there any low-level formatting that needs to
>be done with these RF72 drives, or should they work "out of the
>box"?

Should just work out of the box.


>   Also, just out of curiousity, what does that cable that
>connects to the plug behind the power supply do exactly?  It's
>also in my other BA213, connected to the SDI operator control
>panel, presumably serving the same function.  Did removing
>the plug with the needle give me a good cable, or should
>I go after a proper replacement?

I believe it is bus control. But the real question is, If the cable you 
used was "stock" why was it polarized and the connector not? I haven't got 
a control panel in front of me so I can't tell if it is or isn't.

>One other interesting observation.  Since I was only trying
>to get one drive working for the moment, I only had one
>plugged in.  But the computer will not get past 'F' in the
>patch panel LED without at least two drives plugged in.
>(This indicates power OK).  Needless to say, it took a
>few minutes to figure that out.

BA213 power supplies won't "light up" unless they have enough load on them. 
I've got a dead SCSI drive in one as a 'load' for the left hand supply. 
Without both supplies you won't get DCOK and you'll be stuck at the F display.

Good Luck Doug!
--Chuck