topic Re: OSPF to BGP redistribution in MLX switch in Data Center (MLX & CER/CES)

OSPF to BGP redistribution in MLX switch

Alice — Fri, 06 Mar 2020 02:04:25 GMT

How to configure ospf to bgp redistribution in MLX switch.

Re: OSPF to BGP redistribution in MLX switch

Michael_Morey — Fri, 06 Mar 2020 03:11:31 GMT

Alice,

It is hard to say without knowing more information: Version of NetIron you are running, topology and existing config. The best thing I can point you to is the current L3 Routing guide for our Target Path NI6.2.0

L3 Routing Guide for NI 6.2.0

In general, if you want to redistribute a route from one protocol into another, you would configure this in the Destination protocol ie:

device# configure terminal
device(config)# router bgp
device(config-bgp)# redistribute ospf
device(config-bgp)# redistribute ?
 bgp Border Gateway Protocol (BGP)
 connected Connected
 isis Intermediate System to Intermediate System (IS-IS)
 ospf Open Shortest Path First (OSPF)
 rip Routing Information Protocol (RIP)
 static Static routes

I hope this helps answer your question.

Re: OSPF to BGP redistribution in MLX switch

Alice — Fri, 06 Mar 2020 18:23:52 GMT

Hi Michael,

Here is the topology and the behavior i noticed:

I have loopback interfaces in all the routers. All the loopback interfaces are part of OSPF except in CER-3.

When I redistribute the BGP/RIPv2 routes into OSPF, the BGP/RIPv2 routes are advertised properly to other OSPF routers.
The CER-1 receives a LSA type-5 for the routes 20.0.0.0 and 100.100.100.100 network and able to ping the loopback on CER-3.

I have redistributed the OSPF routes into BGP/RIPv2 on MLX-4 following the config that you shared however, I dont see any
routes on CER-3 for 24.0.0.0 and 34.0.0.0 prefixes. These network doesn't exist on CER-3's OSPF database and routing table.
All the other routes from Area 0 are seen on CER-3.
Not sure what is missing in config but looks like the MLX-4 doesn't redistribute the OSPF routes into BGP or RIP. I have tried
this with both BGP and RIP and the behavior remained same.
The routes 24.0.0.0 and 34.0.0.0 get redistributed only when I configure the redistribution of "connected" routes into BGP or RIP.
Is this an expected behavior?

Re: OSPF to BGP redistribution in MLX switch

Michael_Morey — Fri, 06 Mar 2020 22:13:34 GMT

Alice,

Yes this is the behavior I would expect. While the routes are distributed as part of the OSPF network, from MLX4’s perspective and how it knows these routes, they should simply be Connected routes. You should be able to verify this by looking at MLX4’s routing table, and if they show as type “D” then they are learned as connected and will only redistribute using the “Connected” syntax.

Re: OSPF to BGP redistribution in MLX switch

Alice — Sat, 07 Mar 2020 00:02:42 GMT

Fair enough.

I was comparing the behavior with CISCO lately but I agree with your point.

Thanks a lot.