Troubleshooting

OSPF Neighbour Stuck in EXSTART / EXCHANGE

VSVipul Sir · 6 min readLast updated: 22 July 2026

An OSPF neighbour stuck in EXSTART or EXCHANGE is almost always an MTU mismatch. OSPF advertises interface MTU inside its DBD packets; if the value received is larger than the receiving interface's IP MTU, the router silently drops the packet and the adjacency never reaches FULL.

What you see

%OSPF-5-ADJCHG: Process 1, Nbr 2.2.2.2 on GigabitEthernet0/0 from EXCHANGE to DOWN,
  Neighbor Down: Too many retransmissions

Why it happens

OSPF forms an adjacency in stages: DOWN → INIT → 2-WAY → EXSTART → EXCHANGE → LOADING → FULL. EXSTART is where the two routers elect a master/slave and start exchanging Database Description (DBD) packets. Those DBD packets carry the sending interface's MTU.

If Router A advertises an MTU of 1500 and Router B's interface is set to 1400, Router B rejects the DBD — it cannot accept a packet describing a larger MTU than it can receive. Neither side reports a clean error; the adjacency simply flaps between EXSTART and EXCHANGE, often with retransmission messages.

The usual real-world causes are a tunnel interface (GRE/IPsec lowers MTU), a sub-interface with a manual ip mtu, or one side of a link configured for jumbo frames.

How to confirm it

show ip ospf neighbor
show ip ospf interface GigabitEthernet0/0 | include MTU|Line
show interface GigabitEthernet0/0 | include MTU

Compare the MTU on both routers. show ip ospf neighbor showing a state that never settles on FULL, combined with different MTU values, confirms it.

The fix

! Preferred — make the MTUs match
interface GigabitEthernet0/0
 ip mtu 1500
!
! Only if you genuinely cannot match them (e.g. a tunnel)
interface Tunnel0
 ip ospf mtu-ignore

Matching the MTU is the correct fix. ip ospf mtu-ignore suppresses the check rather than solving it — acceptable on tunnels, a bad habit on normal links, because a genuine MTU mismatch will then cause silent fragmentation problems instead.

How to stop it happening again

Standardise MTU across a routing domain and document any interface that deviates. On tunnels, set MTU explicitly rather than relying on defaults.

Questions people ask

Does ip ospf mtu-ignore fix the underlying problem?

No. It only stops OSPF checking the MTU field, so the adjacency forms. The MTU is still mismatched, which can cause dropped or fragmented traffic later. Match the MTUs unless the link genuinely cannot support it.

Why does the neighbour show FULL on one side and EXSTART on the other?

That is normal for this fault. The router with the larger MTU can accept the smaller DBD packets, so it progresses further than its neighbour, which is dropping the oversized ones.

Reading the fix is not the same as doing it

Every fault on this page can be broken and repaired on real Cisco routers and switches in our Ahmedabad lab — which is how it stops being theory. Come and see the rack before you pay anything.