OSPF flapping

[misaleh]

I am having a problem in my OSPF network. I have a connectivity with my ISP via OSPF that connects us to a remote site but the neighborship keeps flapping. The neighborship is via a sub-interface, there are 3 other subinterfaces on the same physical interface and they don't flap. There are no CRCs on the interface. I've tried changing the hello and dead timers to 30 and 120 respectively, but the issue still occurs. I've also added " ip ospf mtu-ignore", but it also didn't fix the problem. I'm not sure what exactly is the issue or how to fix it so any help would be really appreciated.

The OSPF config. on my router is:

Router ospf 12 vrf WAN
router id x.x.x.x
network 10.0.0.0 0.255.255.255 area 0
default-information originate

interface gi0/0/2.1142
bandwidth 20000
encapsualtion dot1q 1142
ip vrf forwarding WAN
ip address 10.x.x.x 255.255.255.252

and the OSPF configuration on the ISP router is:

router ospf 1
router-id x.x.x.x
redistribute static subnets
network 10.x.x.x 0.0.0.3 area 0

interface Ethernet0.1142
encapsulation dot1Q 1142
ip address 10.x.x.x 255.255.255.252

Here are the logs that appear on the router:

Aug 25 11:51:06: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 11:55:23: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 11:56:25: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 11:59:20: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:02:12: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:03:27: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:06:56: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:10:47: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from FULL to DOWN, Neighbor Down: Dead timer expired
Aug 25 12:12:30: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:14:21: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:16:47: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done
Aug 25 12:20:05: %OSPF-5-ADJCHG: Process 12, Nbr 10.x.x.x on GigabitEthernet0/0/2.1142 from LOADING to FULL, Loading Done

[paul driver]

Hello

---

@misaleh wrote:

Will try it tomorrow, but what if I want to advertise other networks 10.x.x.x in OSPF?

---

you will be - ospf is attached to the interface 10.x.x.x is assinged to 

int x/x
ip ospf x area x

As for the debug -- 

---

@misaleh wrote:
Cannot see ourself in hello from 10.200.14.14

---

Check to make sure the neighbors can ping each other correctly - perform an extended ping to see it fails, check your interfaces statistics for crc errors.

---

[Richard Burts]

Paul notes the debug message about not see ourselves in hello and suggests a possible issue with IP connectivity. I am convinced that this is not a problem but is normal expected behavior if the interface on the neighbor has just come up from a down state. The interface that just came up sends an OSPF hello and of course there are no neighbor known at this point so the hello has no neighbor information. Looking at the debug we see the various stages of negotiating the neighbor relationship and culminates with a successful full relationship

Aug 25 08:42:21.293: OSPF-12 ADJ Gi0/0/2.1142: Synchronized with 10.200.14.14, state FULL

The original poster asks "what if I want to advertise other networks 10.x.x.x in OSPF?". Perhaps there is something in this question that I do not yet understand, but it seems to be a simple question, and I answered it in a previous post. On the customer router the OSPF network statement with the /8 mask will result in any interface whose IP address is in the 10.x.x.x network will run OSPF and that subnet will be advertised. On the ISP router with the /30 mask it wold require an additional OSPF network statement (or configuring OSPF directly on the interface as Paul suggests).