| Title: | FDDI - The Next Generation |
| Moderator: | NETCAD::STEFANI |
| Created: | Thu Apr 27 1989 |
| Last Modified: | Thu Jun 05 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2259 |
| Total number of notes: | 8590 |
Hi, I was wondering if anyone could help me with the following problem. I have two DECbridge 620's set up in a dual ring configuration, serving three ethernet segments. One bridge is the live bridge, the other bridge is the backup bridge, the three ethernet ports on the live bridge are in Forwarding mode, the three ports on the backup bridge are in Backup mode, as you would expect. Some time ago, I switched over what was then the live bridge to be the backup bridge, and the backup bridge to be the live bridge, so I know they are both capable of forwarding onto the three segments okay, that's a bit of background information for later. On Saturday, a problem developed whereby both the live bridge and the backup bridge were forwarding onto one of the segments at the same time. This was only happening on one of the three segments, the other two were fine. I know this was happening because the FWD light was on on both bridges, and the live bridge , which is also the root bridge, was clocking up Bad Hello Limit Exceeded messages. I had a PC set to monitor the segment and could see the other Hello's coming from the backup bridge. I reckoned there was probably a problem with the backup bridge, so I removed it from the dual ring completely, thinking I could just run with one bridge until I could get the faulty one fixed/replaced. Anyway, what happened was that when I removed the backup bridge from the dual ring, the segment in question filled up with Framing and CRC errors. The PC monitor I had set up showed this, and the line counter on the remaining bridge also showed it. Almost every system on the segment was transmitting bad frames. The only way I could stop the errors on the segment was to connect in a new, working DECbridge back into the dual ring again. This meant I was back in the situation of having a live bridge and a backup bridge again. I can live with the fact that one of my bridges failed, that's okay, it's the reason I had a live and a backup. What I can't understand is why the segment in question would have so many errors when I only had one bridge connected. The other two segments served by these bridges were okay, and why would the problem disappear when I connected another bridge back in again. Would anyone like to comment? Jim.
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 967.1 | KONING::KONING | Paul Koning, A-13683 | Tue May 25 1993 15:54 | 8 | |
If the Ethernet transmitter goes bad, you'd get all those bad frames. If among those are bad hellos, the other bridge on that LAN would end up going into forwarding because it doesn't see any good hellos from a higher priority bridge. Since the three Ethernet ports each have their own Ethernet chips, it certainly makes sense for one port to misbehave while the others work fine. paul | |||||
| 967.2 | Which bridge to blame? | BUSSTP::JHANNAH | Wed May 26 1993 08:58 | 13 | |
My thinking was, since my "backup" bridge is forwarding when it should
be in backup mode, then there's probably something wrong with it, so I
swapped it for a good one. I suppose it could have been the "live"
bridge which was at fault. If it's ethernet port was faulty then that
could explain why the backup bridge kicked in, it would also explain
why I had framing errors when I only had one bridge connected, the
"live" one. I still have the "live" bridge connected in and serving
the segments just now. If it was the problem, and could still be since
it's there just now, and it was the cause of the framing errors, why
would they disappear when I reconnected in a backup bridge?
Jim.
| |||||
| 967.3 | Backup Configs not Always 100% Wonderful | ROCKS::CAMP | Tue Jun 01 1993 08:10 | 24 | |
Interesting problem and it may be that as Paul said if the "main"
bridge was sending bad packets then the bridge hello messages are
likely to have errors so the backup bridge can't see theses hellos,
so its starts to forward. Now the main bridge starts to see these
hello messages, which it believes are incorrect (assumes the "main"
bridge can receive OK), and the bad hello counter will clock up
and I assume that port will then stop transmitting, attempt a
reinitialize on that port, and come back on line listening, see the
hello's from the backup bridge, and reinit etc....
So it's likely that the main bridge never gets back on line, hence you
don't see the bad packets. The GOTCHA here is that the Forward light
should not come on on the main bridge.
This is the Catch 22 with backup bridge configurations is that if either
the main or backup bridge dosen't fail "nicely", eg PSU blows up, internal
logic gets a fatal error etc, but fails marginally then it can really
screw up the network, and more so if the bridge is or is near the route
bridge. If the failed bridge can't detect that its gone wrong its
bad news. Not that common I'm pleased to say but it can happen.
(Some Ethernet products on the market can't perform a self test on to the
network in any event, ie they can transmit onto the cable and receive
from it, but can't truly receive their own transmitted message.)
| |||||
| 967.4 | Works, but need to test more. | BUSSTP::JHANNAH | Tue Jun 01 1993 12:18 | 10 | |
Everything has been running okay for the last week and a half, I'm
pleased to say. What I'm going to do is, now that I've changed the old
"backup" bridge for a new one, test the failover to make sure that it
works okay, and also disconnect the "backup" bridge to make sure that I
can run with only the "live" bridge connected into the fibre ring.
I'll post the results when complete.
Jim.
| |||||