[Info-vax] Trying to boot a MicroVAX II with VMS 7.3

[johnwallace4 at yahoo.co.uk]

On Mar 29, 9:43 am, "Robert Jarratt" <nos... at nosp.am> wrote:
> "Robert Jarratt" <nos... at nosp.am> wrote in message
>
> news:Lwxzl.125722$C87.27817 at newsfe30.ams2...
>
>
>
>
>
> > "Richard B. Gilbert" <rgilber... at comcast.net> wrote in message
> >news:P9adnSydZrhJ91PUnZ2dnUVZ_jaWnZ2d at giganews.com...
> >> johnwalla... at yahoo.co.uk wrote:
> >>> On Mar 28, 2:44 pm, "Robert Jarratt" <nos... at nosp.am> wrote:
> >>>> I am trying various ways to boot my MicroVAX II. I am currently unable
> >>>> to do
> >>>> this using MOP because 7.3 does not support DEQNA. So I am now trying
> >>>> standalone backup using TK50. I created the kit on another machine
> >>>> running
> >>>> 7.3 and put it into the MicroVAX II. After a while I got:
>
> >>>> %SYSBOOT-F-PFN allocation overwrites CI ucode.
>
> >>>> This seems to be related to SYSGEN parameters essentially requesting
> >>>> too
> >>>> much memory. Is there a way to generate a standalone backup kit on
> >>>> another
> >>>> machine such that I can avoid this issue?
>
> >>>> Thanks
>
> >>>> Rob
>
> >>> [interim answer till the usual - more reliable - oldtimers turn up]
>
> >>> The ideal solution (since your VAX is actually networked) might be to
> >>> find a DELQA from somewhere and junk the DEQNA, so that you are not
> >>> later caught out by the DEQNA lockup "quality improvement
> >>> opportunity" (it's not a problem, it's not an issue, it's a QIO). They
> >>> might be expensive for a hobbyist machine though.
>
> >>> Why is a MicroVAX worrying about CI microcode anyway?Is there a CI
> >>> interface in this MicroVAX (it's unlikely but not impossible in a Qbus
> >>> system)? Or is the error message misleading?  Did that error message
> >>> get printed as you seem to say, or did you look it up based on a
> >>> numeric error status value? If numeric, are you confident you got the
> >>> decimal/hex thing right (I can't remember the details or check right
> >>> now but I can remember being caught out by the difference between $
> >>> exit 910 and $exit %X910 etc).
>
> >>> I have a very vague recollection that the DEQNA check could at one
> >>> time be overridden by boot flags but can't currently find any evidence
> >>> of this (and would be very surprised if it lasted as late as VMS 7.3
> >>> anyway).
>
> >>> If you are worried about SYSGEN parameters, have you tried booting
> >>> conversationally and then the usual conversational boot options are
> >>> available to you?
>
> >>> Whatever happened to VAXrsm (Remote System Manager) anyway... netboot
> >>> a VAX and install VMS over the LAN without actually needing to be
> >>> clustered (plus many other useful things which I have mostly
> >>> forgotten). It might even not have a DEQNA check in it.
>
> >>> Good luck.
>
> >> I don't think I've seen or heard of RSM in the last ten years or so!  It
> >> may have become a casualty of the acquisition by Compaq and then HP.  I
> >> used to use it at McGraw-Hill to handle the VAX Clusters at about 100
> >> field offices.
>
> > The error message is exactly as printed. I don't know what CI interface
> > is, can someone tell me?
>
> > I would love to get hold of a DELQA, but eBay prices are, as you surmise,
> > prohibitive for a hobbyist. I am looking instead at getting hold of an
> > older version of VMS.
>
> > Regards
>
> > Rob
>
> I may have found a possible cause of my booting problems. I decided to
> remove everything except the CPU, the tape controller and the DEQNA, so that
> included removing the two memory boards (which incidentally look like they
> are 4MB each). When I did this the console software did not come up when I
> switched on the machine. The CPU should have 1MB on board, so I expected it
> to be able to run the console software if nothing else. Have I got that
> wrong? Should I expect it to show some signs of life with only the onboard
> memory?
>
> Regards
>
> Rob

Taking out things you don't 100% trust is A Good Idea but you may have
gone a bit too far by taking out both the memory cards. 1MB of user
memory on the CPU, and signs of life with just that 1MB without
external memory? Are you really sure? 1MB on board would likely mean
that typical MicroVAXes with additional memory would have odd (5, 9,
17 etc) memory sizes  ?

The CI was the original "Cluster Interconnect" typically used to
connect between cluster members and an HSC50/70 or similar storage
controller (gross oversimplification). Typically the interfaces are
not seen on Qbus systems, although iirc a third party one was
available. Hence the comment/question - why is it trying to load the
CI microcode (maybe this is expected, maybe it isn't).

The VMS docs have a section on troubleshooting cluster-satellite
booting; it's probably worth having a read there if you haven't done
so already, just in case. E.g.
http://h71000.www7.hp.com/doc/731final/4477/4477pro_024.html#satellite_trouble_sect

As a different test of how well network booting is working, have you
tried
B/100 READ_ADDR
which (in the right circumstances, see the cluster troubleshooting
book) should netboot a simple program which reads the MAC address of
the network adapter. Might be BOOT /R5:100 or some similar variant,
depending on console version. It'll probably work fine or you wouldn't
have got as far as you already have.

I'd probably put back the two memory cards, one at a time, and see
what happens (again, if you haven't already done that).