One of the most popular questions is: Why is the ip ospf mtu-ignore command considered dangerous?
An MTU mismatch between two OSPF routers causes the adjacency process to fail during the exchange of Database Description (DBD) packets, preventing synchronization of the Link-State Database (LSDB). The ip ospf mtu-ignore command is introduced used to bypass this issue by forcing both routers to ignore the MTU verification during the Exchange state or DBD packets exchange.
The figure illustrate what exactly happen when the ip ospf mtu-ignore command is enabled between two adjacent OSPF routers with different MTU value.
First The crucial step is when a largeLink State Update LSU packet is generated in the loading state as the response to the Link State Request packet.
Let’s assume router R1 has an MTU of 1500 bytes, and router R2 has an MTU of 2000 bytes, while all other routers in the OSPF domain are using the default MTU of 1500.
Now, imagine R5 boots up and begins advertising 500 new networks. It sends a Link-State Update (LSU) packet containing its Type-1 LSA, the R5's LSA Type-1 is listing: 500 and 2 stub networks (inter-router links). R2 replicates and forwards the LSU packet to R1 through it's interface with MTU 2000.
The LSU packet, now large due to the size of R5's Type-1 LSA, exceeds 1500 bytes. Since R1's interface has a lower MTU, it drops or fails to process the LSU packet.
Finally, R1 is unable to synchronize its Link-State Database with R5's new Type-1 LSA and cannot install routes to the newly advertised networks."
Because the MTU mismatch was ignored earlier (due to the command), the routers can proceed with the synchronization but the problem occurs when the router with higher MTU value is replicating or generating a large LSU packet in the loading state that exceeds the MTU of the neighbor router.
