Incurring OSPF Issues: Dead Timer Expired, Too Many Retransmissions

[carlton_skanska]

Hello Community,

Can someone please help shed some light on why we're getting so many OSPF errors surrounding dead timers, Too many retransmissions, and Ignore timer expired?

The errors appear on our Tunnel 0 interfaces.

I have checked the Tunnel neighbour interfaces for any mismatches and everything appears to be ok.

I have attached the configs.

This is a rather major issue here at work, so your expeditious help will be greatly appreciated.

Cheers

Carlton

[carlton_skanska]

Hello Community,

Here are a few typical show logg messages on the hub router:

8 01:12:42.016: %OSPF-5-ADJCHG: Process 1, Nbr 10.45.121.3 on Tunnel0 from EXCHANGE to DOWN, Neighbor Down: Too many retransmissions

May 28 01:13:42.017: %OSPF-5-ADJCHG: Process 1, Nbr 10.45.121.3 on Tunnel0 from DOWN to DOWN, Neighbor Down: Ignore timer expired

May 28 01:15:45.854: %OSPF-5-ADJCHG: Process 1, Nbr 10.45.121.3 on Tunnel0 from EXCHANGE to DOWN, Neighbor Down: Too many retransmissions

May 28 01:16:45.855: %OSPF-5-ADJCHG: Process 1, Nbr 10.45.121.3 on Tunnel0 from DOWN to DOWN, Neighbor Down: Ignore timer expired

May 28 01:28:37.082: %OSPF-5-ADJCHG: Process 1, Nbr 10.45.121.3 on Tunnel0 from EXCHANGE to DOWN, Neighbor Down: Too many retransmissions

Cheers

Carlton

[Steve Fuller]

Hi Carlton,

Is this a new branch site you're just adding or has it worked in the past and suddenly broken?

The reason I ask is that OSPF never seems to make it past the Exchange state, which is often caused by MTU mismatch between the two routers. This problem and the debug to run to confirm are described in the Why Are OSPF Neighbors Stuck in Exstart/Exchange State? technote.

Regards

[carlton_skanska]

Hi Steve,

Thanks for responding.

This site is a live site and working fine as we speak. However, there are certain times of the day where users complain of sluggish (or no internet) network. When I check the router logs I see the dead timer messages etc...

I'm also leaning towards the problem being an MTU issue.

It has been suggested on this site that a fix would be to implement ip ospf ignore-mtu on the spoke interface.

What do yout think?

Cheers

[Steve Fuller]

Hi Carlton,

I've never used the ip ospf mtu-ignore command, but it looks as if it could be an option.

I guess you could be running into a situation where the path the tunnel is established via changes, with that path having some lower MTU than the path used when the site is working OK.

It might be worth trying to capture ping and traceroute with large packet size and DF bit set when the branch is working OK and when it fails to check if the MTU is lower when it has the problem.

Looking at the timestamps on the logs it could be pretty unsociable to get that though

Regards

[paul driver]

Hello Carlton,

I also see you have CBAC enabled, on incoming_from_internet acl's - is ospf being permitted without being subjected to

the CBAC ruling?

res

Paul

Please don't forget to rate any posts that have been helpful.

Thanks.

[carlton_skanska]

pdriver,

Sorry for late response. I believe its an MTU issue. We're going to keep on investigating.

Thanks again for responding...

Cheers

[carlton_skanska]

pdriver,

Are you able to see the my ratings?