T.R | Title | User | Personal Name | Date | Lines |
---|
1205.1 | | CRONIC::LEMONS | And we thank you for your support. | Mon Jul 08 1991 12:33 | 5 |
| Is this really that hard, or am I so totally off-base that no response is
possible?
Thanks!
tl
|
1205.2 | One idea.. | SUBWAY::REILLY | Mike Reilly - New York Bank District | Mon Jul 08 1991 13:29 | 43 |
| This is not so easy. Multi segment LANS normally don't have a
static configuration. If you determine the path between your load
host and the target system and store it in a file, there is a very
good chance that this is not the path which will be in use when you
need it.
Any tool you develop will have to determine the current bridge
configuration before it attempts to change the filters.
Assuming you are only using DEC bridges on your LAN, the following
procedure may work.
Power up the target machine and let it issue a load request. This
will place it's address in the bridge forwarding tables. You stated
that ALL your bridges filter out MOP requests. If this is so then
only bridges attached to the segment that the target machine resided
on will hear the MOP request and add the address to it's forwarding
tables. Using ELMS or DECmcc you can query each bridge on your LAN
and ask if the bridge has your target machine in it's forwarding
table. Bridges that know about the target machine are attached to
the same segment.
With this info you now known the name of a bridge attached
to the segment the target machine resides on.
Trace the path from this bridge up to the root bridge. This is
accomplished by asking the bridge to show it's root bridge and it's
designated bridge. It's designated bridge is the next step on the
way to the root bridge. Check both line 1 and line 2 of each bridge
as only one of the lines will be pointing to the real designated
bridge. You will now have a list of bridges which connect the target
segment with the root segment.
Perform the same trace from the segment attached to the InfoServer.
Compare the two lists of bridges and locate the first bridge common
to both lists. The bridges between this common bridge and the end
segments must have the MOP filter removed for a down-line load to
work.
- Mike
|
1205.3 | | EISNCG::OLEARY | | Mon Jul 15 1991 16:44 | 11 |
|
I may be wrong, but I thought a network Assets tool known as POPENIM
did a function "similar" to what you are looking for. From what I've
been described, POPENIM used RBMS data to correctly place nodes on the
NMCC/VAX ETHERnim map (database). I do know that POPENIM no longer
works with some of the later releases, but you might be able to get
some hints/ideas from the code.
Network assets tools are discussed in the HORUS::NASSETS conference.
Mike
|
1205.4 | Light dawns . . . | CRONIC::LEMONS | And we thank you for your support. | Fri Jul 26 1991 18:38 | 13 |
| Mike
Thanks again for your suggestions in .2! I'm finally writing the procedure
to accomplish .0, and I have a question:
> Check both line 1 and line 2 of each bridge as only one of the lines will be
> pointing to the real designated bridge.
And how can I tell which line points to the real designated bridge? By choosing
the line with the lowest Root Path Cost? Or is there another way?
Thanks!
tl
|
1205.5 | I see a monster DCL procedure brewing.. | SUBWAY::REILLY | Mike Reilly - New York Bank District | Mon Jul 29 1991 22:04 | 7 |
| Check both of the lines as the designated bridge for one of the
lines will be the bridge itself. Be careful with LB200's as they
have two addresses and either address can be returned as the
designated bridge.
- Mike
|
1205.6 | Naw, no monster. | CRONIC::LEMONS | And we thank you for your support. | Tue Jul 30 1991 22:31 | 15 |
| >Check both of the lines as the designated bridge for one of the
>lines will be the bridge itself.
You're right (I checked). One line showed a designated bridge of the
real designated bridge; the other line showed itself as the designated
bridge.
>Be careful with LB200's as they
>have two addresses and either address can be returned as the
>designated bridge.
Sorry, I'm not sure I understand your point.
Thanks!
tl
|
1205.7 | Bridge Funkiness | RACER::DAVE | Attending The School of Comparative Irrevelevance | Wed Jul 31 1991 09:47 | 22 |
|
LB200's use two different address ROM's, one for each port.
In the normal case, you will see two addresses that are sequential,
e.g.
08-00-2b-13-de-4b and 08--00-2b-13-de-4c
Note that there is NOTHING that enforces this, it is only a quirk
of the manufacturing line procedures that this is true. One could
just as well see a LB200 that has completely different address that
appeared to be totally random.
DECbridge-500's use the "shadow" address to give the "other" port
its address. E.g. the address of one port will be, for example,
08-00-2b-15-e4-5f, and the other address will be that address ORed with
%X'40-00-00', i.e. 08-00-2b-55-e4-5f. (The shadow address is on the Ethernet
side.)
I am not sure what the new FDDI to Ether bridge will do with its three
port cards, as I have not noticed on on the net here in LKG. I know there
is one, however, and I guess I should go figure out what it does for
addresses.
|
1205.8 | | SUBWAY::REILLY | Mike Reilly - New York Bank District | Wed Jul 31 1991 10:03 | 6 |
| A LB200 has a different ethernet address on each line. If the
designated bridge for the device your are examining is a LB200 the
designated bridge address returned can be either of the LB200's addresses.
- Mike
|
1205.9 | | CADSYS::LEMONS | And we thank you for your support. | Mon Aug 17 1992 17:50 | 13 |
| Re .2
Good advice. My situation is additionally complicated in two ways:
1) the ~ 40 bridges on my LAN may or may not be filtering MOP at the time I
run my procedure, and
2) it would be nice to have the procedure work on nodes that are running, or
were just shutdown (a running node may have been known to many bridges, depending
on which noes it was communicating with).
Other thoughts? Has anyone written a tool like this already?
tl
|
1205.10 | sanity check needed | CADSYS::LEMONS | And we thank you for your support. | Wed Aug 19 1992 14:10 | 14 |
| Let's say I do the following:
MCC> show bridge * for dat dyn ent <source-address> outbound port, to file a.a
MCC> show bridge * for dat dyn ent <target-address> outbound port, to file b.b
Then I compare these two files. Where I find a bridge that hears the source and
target on different ports, I know that this bridge stands between the two. For
my purposes, I then set this bridge or these bridges to pass MOP, and I'm off
and flying.
Or am I missing something here?
Thanks!
tl
|
1205.11 | | CADSYS::LEMONS | And we thank you for your support. | Tue Sep 01 1992 15:13 | 1 |
| hello?
|
1205.12 | | SLINK::CHILDS | The Fringe | Wed Sep 02 1992 11:37 | 9 |
| Hello. :-)
I'm not sure that the folks that can best answer your questions about
this are here reading this file. You might want to ask in
NAC::RBMS_LANBRIDGE100. That conferences is more specific to the inner
workings of bridges, this conference is mostly bridge management as it
relates to DECmcc.
Press <Select> or <KP7> to add to your notebook...
|