08-01-2012 05:26 AM
I have some older Jetcore switches that we have in production. These are receiving a "filtered" internet feed. They currently peer with 5 other BGP peers.
They have two uplink providers within this mix. I have found that if I have any link changes or even a bgp reset on one of the uplinks that the jetcore switches would go into and stay in a bgp active stage and never to to established. After a lot of debugging i found that the router they were connecting too was getting these debug messages as if another bgp peer was trying to take over the peering session. This other bgp peer is a much newer MLX switch. I then noticed that the only jetcore peering sessions that had this issue were connected via a VE interface all the others that were in use were stable and were using the physical ethernet interfaces.
So i reconfigured the vlan and used a physical interface for the two switches having the problem and it has fixed the issue!!
So my question is, has anyone else seen this behavior? Is there a known bug on these older switches, using the ve interface?
Advice is helpful!
08-03-2012 07:47 AM
Yes, I took the ip address off the VE interface and then put it on a physical interface and it fixed the problem. The bgp peering session came right up. Yes we have 3 of these jetcore switches in production. We just upgraded to them 2 1/2 years ago..Dont laugh! I know they are old..
08-03-2012 08:00 AM
Sorry I have not seen this. if you have support, I would check the release notes on the code train you are on to see if that is indeed a bug or not.