| Hi Sam,
Below is a response to some (but not all) of the questions you asked.
>
> MCC sends out the loopback command, the DEMMR responds
> correctly (we monitored on LAN analyzer)
>
There is a major problem with MCC-ETHERNET AM T0.1.0, documented as
MCC_INT QAR#22 (also Note #41 in this notes conference). What it basically
says is that, using the TEST directive through an ethernet port running
DECnet, the BROADCAST ADDRESS (FF-FF-FF-FF-FF-FF) will get buried in the
packet as the Return Address. This means two things:
1. The TEST directive in IFT is basically useless
2. The target Ethernet Station will get a loop packet with
a bogus return address (broadcast)
3. The packet that the Ethernet Station generates will not be
"received" by MCC (ie: "cannot communicate with target")
I am assuming that you didn't scrutinize the packets closely enough to
catch this problem. [Note that the latest ENET AM we gave you does not
have this problem.]
>
> Second, "SHOW station <addr> all char, via port xqa0:", MCC sends out req
> ID command, the DEMMR responds back, but MCC crashes (return back to $)
> with the "improperly handled condition, image exit forced" message.
>
We've tested/verified that the MCC-ETHERNET AM T0.1.0 code handles
*correctly formatted* SYSID responses. Your problem points to a potential
"lack of robustness" (ie: bug :-) in the code being used to parse the
SYSID. Could you send us a dump of the SYSID packet you're sending out
so that we can at least "desk check" it (even if the format is incorrect,
this packet would help us to improve our code)? [Note that the latest
ENET AM we gave you allows this packet tracing to be done from the AM.]
>
> Third, "show station <addr> all counters, via port <port>", MCC sends out
> correct V3 req counters message, but we do not support V3 req counters
> . So it is OK.
>
I'm glad SOMETHING is working right :-)
>
> Which version of MCC will support MOP version 4, IEEE 802 test,xid.
>
The EFT version will support the above requests. [Note that the latest
ENET AM we gave you supports this]
>
> I follow the field test version MCC document, but the register command is
> not working. It seems to me something wrong with the command format.
>
The IFT (T0.1.0) Release Notes should state (he says, hoping that they
actually do...) that REGISTER is not supported for Ethernet Station for
the IFT version.
[Note that the latest ENET AM we gave you does support REGISTER]
>
> If we have problems, we are the MCC field test site (register through Peter
> Savage), any body can help? I don't have QAR to report error so I have to
> use MCC notes file.
>
I would suggest that you:
1. Address any MCC-ETHERNET AM questions on the NEW STUFF WE GAVE YOU to
Wayne (as I believe you are already doing).
2. Continue to use this Notes Conference to report any problems you find
with T0.1.0.
-----------------------------
Please Note:
Two of your questions remain unanswered:
1. General DNS support questions
2. "RAB, MCC initialization failure" problem.
I trust that others in MCC (a MCC$DNS person and a MCC$Kernel person) will
answer those questions for you (hint. hint.).
Chris Brienen
MCC-Ethernet based AMs.
|
|
The DNS problems are being worked off line with Sam Shen.
For anyone experiencing problems finding out how to connect to the
corporate namespace, the namespace nickname is DEC: and each major site
has a nameserver that serves this namespace.
Ask your system manager for the name of the local nameserver. If he/she
does not know, have her/him find out. HINT: mail to GLUTNY::SYSTEM
asking the name of the server associated with your site might result in
success.
The documentation on the DNS nameserver is too voluminous to include in
the MCC installation documentation.
A complete introductory explanation is included in the VMS 5.3 New
Features manual.
...kjn
|
|
>>sometimes I run MCC and got can not connect RAB, MCC initialization failure
>>, and I can't get MCC prompt< what is it?
This error message has never come up before, at least that I know of.
I'm assuming that there are at least two error messages, the first, can
only be an RMS error message, since there are no MCC error message
which mention RABs (at least that I know of). The second one on
initialization failure appears to be coming from the
Enrollment/Dispatcher portion of the Kernel, in which as you should
have gotten something like
%RMS-F-CCR, Cannot connect RAB
...
%MCC-F-KERNINITF, MCC initialization failure
If this is true, then it appears as though one or more of the MCC
system files, has been corrupted or moved to another location. Are you
using DECNET or DFS to access the MCC system files?
-Matt.
|