T.R | Title | User | Personal Name | Date | Lines |
---|
1510.1 | | NETCAD::STEFANI | Look Ma, I'm drinking my oatmeal! | Wed Nov 30 1994 09:44 | 42 |
| Brad,
I don't think your setup will work. DC1 and DC2 are connected fine and
the Upper Tier is a normal dual-ring configuration. However, DC3 and
DC4 are not being dual-homed properly. I would change the layout to:
Upper Tier
_____________________
| _________________ |
| | | |
---------- ----------
|A B | |A B |
| DC1 | | DC2 |
|M M | |M M |
---------- ----------
| |_____________ | |
| | | |
| ________________| |
| | Lower Tier | |
| | |__ |
| | | |
---------- ----------
|A B | |A B |
| DC3 | | DC4 |
| M M | | M M |
---------- ----------
| | | |
| | ------------ |
| --|----------- |
| | | |
---------- ----------
| A B | | A B |
| DECnis | | Vax |
---------- ----------
I realize that the customer may want to limit the fiber strands going
from the Lower Tier to the Upper Tier, but the above configuration
should work and provide the fault tolerance they're looking for.
- Larry
|
1510.2 | | CSC32::B_GOODWIN | MCI Mission Critical Support Team | Wed Nov 30 1994 11:23 | 8 |
| Thanks Larry,
I was pretty sure my original config was wrong and I would have to configure it
the way you did. But the customer wanted to save running the two extra fiber
links, because he is running them in diverse routes to avoid having a possible
fiber cut on all four fiber links at the same time.
Brad
|
1510.3 | | NETCAD::STEFANI | Look Ma, I'm drinking my oatmeal! | Wed Nov 30 1994 11:52 | 11 |
| >>links, because he is running them in diverse routes to avoid having a possible
>>fiber cut on all four fiber links at the same time.
That makes a lot of sense. In our NetWare SFT III MSL dual-attach
configurations, we recommend that each cable is placed in separate
conduits.
For your customer, I suggest at least running the two cables from DC1
in one conduit and the two cables from DC2 in another.
- Larry
|
1510.4 | another way | NETCAD::ARGO | | Thu Dec 01 1994 17:06 | 55 |
| Hi Brad.
>The B port on DC3 and the A port on DC4 led status shows that they are in
>Standby/Dual Homed. If I pull either M port connection on DC1 or DC2, the
>leds on DC3 B port and DC4 A port stay in the Standby/Dual Homed.
>Effectivly isolating the DC that the M port was pulled on. Is this normal
>behavour?
What you have done is to present concentrators DC3 and DC4 with
a tree connection (A or B to an M) and a peer connection
(A to B). The topology rules say to take the tree connection
and reject the peer connection. If you pull the DC3's A-M
connection, DC4 is going to continue to reject the A-B connection
because is it is still presented with a choice between a
tree connection and a peer connection.
If you ignore the religion about dual homing verses putting
it in the dual ring, you can do something like the following
to avoid running the extra cable. It depends on where you
think the network is most at risk.
Upper Tier
+----+ _________________ +------------+
| | | | | |
| ---------- ---------- |
| |A B | |A B | |
| | DC1 | | DC2 | |
| | | | | |
| ---------- ---------- |
| |
+------------------------+ |
| |
+-------------------|------------+
| +---------------+ |
| | Lower Tier | |
---------- ----------
|A B | |A B |
| DC3 | | DC4 |
| M M | | M M |
---------- ----------
| | | |
| | ------------ |
| --|----------- |
| | | |
---------- ----------
| A B | | A B |
| DECnis | | Vax |
---------- ----------
Dennis
|