[Info-vax] Reject Cause=Invalid Message Format, Erroneous Transport PDU='386200BD00000005'H

[Johnny Billquist]

On 2016-11-27 03:07, Dirk Munk wrote:
> Scott Dorsey wrote:
>> Supratim Sanyal wrote:
>>> Hi,
>>>
>>> I am no network expert and looking for help.
>>>
>>> Two SIMH OpenVMS 7.3 DECnet Plus nodes - QCOCAL (1.550) and CLOUDY
>>> (1.551), connected over Johnny's UDP bridge. FAL directories across the
>>> network work fine. However I cannot get NCP status from the remote host
>>> from either. Also the circuits do not seem to be setting themselves up,
>>> I don't know if that is a cause or effect of the Remote Protocol Errors.
>>>
>>> Note: A SIMH PDP-11/24 (JUICHI, 1.552) runnign RSX-11M-PLUS is colocated
>>> (running on the same linux host) with CLOUDY, and NCP queries somehow
>>> work fine to it.
>>
>> RSX-11+ supports Phase V?
>> I don't think it does.
>>
>> So, if the query is working on the RSX-11 node, it is likely a Phase IV
>> query.  And likely won't work on the Phase V node.
>>
>> Mixing Phase V and Phase IV can be done but is best avoided.
>> --scott
>>
>
> He didn't say that the RSX node is running Phase V, it clearly is
> running Phase IV.
>
> What he is showing is this:
>
> You can use NCP on a Phase V node to get information from a Phase IV node.
>
> You can use NCP on a Phase V node to get some (rather bogus) Phase IV
> information of that same node. Phase V information is somehow
> 'translated' into Phase IV information.
>
> You can not use NCP to get similar 'bogus' Phase IV information from
> another Phase V node.
>
> I'm not surprised.
>
> DECnet information of a Phase V node should be retrieved with NCL, not
> NCP. The same applies to Phase V information from another Phase V node.

People are confusing things.

Who cares if it is NCP or NCL. The protocol used to manage this 
information is called NICE (Network Information and Command Exchange or 
something like that). NCP or NCL are just a user interface to the 
underlying interfaces and protocols.

That said, the Phase V NICE implementation is incompatible with the 
Phase IV NICE implementation (if phase V even have a NICE listener?). 
However, the ways things are incompatible also seem to extend to how 
they give errors back over DECnet, causing different errors in VMS and RSX.

In VMS:
NCP>tell qcocal sho exec
%NCP-F-CONNEC, unable to connect to listener
-SYSTEM-F-REJECT, connect to network object rejected
NCP>

In RSX:
NCP>tell qcocal sho exec
NCP -- Show failed, oversized Management command message
NCP>

Bottom line: Phase V only have limited phase IV compatibility. But we 
all knew that (well, I guess Supratim didn't).

	JOhnny

-- 
Johnny Billquist                  || "I'm on a bus
                                   ||  on a psychedelic trip
email: bqt at softjar.se             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol