pinging over frame-relay to OSPF area0

[cymurai]

I am a Cisco Networking Academy student working on my CCNP. I am finishing a case study on Remote Access.

I seem to have a puzzling problem that even my instructor is stumped.

Scenario: I have 3 core routers running OSPF (area0), all other remote routers have static default routes to their respective core router. All core routers have redistribute static subnets. One core router is connected to its remote router by frame-relay (using Adtran as cloud); I am not using sub-interfaces. ALL of the other routers can ping the remoteFR router, but the remoteFR router can only ping past the coreFR router using an extended ping using the loopback address as the source address. It is unable to use standard ping or an extended ping using the outgoing serial port address. (It is however able to ping the coreFRs fa and serial interface) ??????

(Traceroute gives the same results as ping)

After researching for several days, I asked my instructor for help. He got out the instructors configs, but we did not see any major differences. Then he got out the configs he got from his instructor at Cisco, those configs showed that the coreFR router had the frame-relay subnet included in the OSPF (area0). This does seem to fix the problem.

I can see this making the coreFR router an internal router instead of an ABR, which the other two core routers are, but it baffles me as to WHY this makes a difference between being able to ping from the remoteFRs loopback address and pinging from the outbound serial interface address. I have not been able to find any information relating to this problem.

Thank you for any input you may have.

[Erick Bergquist]

This is a routing issue. Without the frame relay subnet in OSPF or redistributed into OSPF the rest of the routers won't know how to get the IP subnet for the remote FR network where ping is sourcing from. When you ping from the remoteFR loopback interface it responds because the core has a static route to that network that is redistributed into OSPF thus other routers know how to get to it.

Hope this helps, erick

[cymurai]

First, thank you for spending the time/effort in trying to help me understand this problem.

OK, I see your point, but then WHY don’t ALL of the other subnets on the other core routers need to be included in the OSPF(area0)? The other two core routers: one has a remoteDialup router using an asynchronous dialup modem link, and the other has a serial link to a routerISP, and an ISDN PRI interface (using dialer profiles) going to two remote ISDN BRI routers, it also provides for ISDN BRI backup for the remoteFR router. ALL of these other routers do not seem to have any problems pinging any other router – even without the coreFR router including the frame-relay subnet, they could still ping the remoteFR router.

It seems that OSPF is treating the FR subnet differently than the other subnets. Only the remoteFR router experiences the problem of pinging from its outgoing frame-relay interface. Using the loopback address it has no problem pinging ALL of the other routers.

[rjackson]

I dont follow the whole picture here but heres a couple of things to think about.

When you ping from a router the source address it uses is the address on the interface the ping goes out. Go to the router/s that do not answer and see if they have routes to that network. When your core router was not in area0 where was it? How did it know where any networks were?

[cymurai]

All core routers have static routes to the remote Lo addresses, and all remote routers have default routes to their core router. Everything seems to work fine when FR-int on core router is included in OSPF area.

If the FR-int is not included, then the only problem is being able to ping from the outgoing FR-interface on the remote router; however extended pings from Lo add work fine???

None of the other core routers seem to need their WAN interfaces included in OSPF. And ALL other remote routers seem to ping everything including the FRremote router.