OSPF - Received invalid packet: Bad Checksum

[shane-cross]

I have been working on a network migration from one frame-relay carrier(MCI) to a new one(Intermedia). The client I am working for has never used any system logging for their WAN. They have 180 remote sites composed of 1600,1700 and 2610 routers for access. The host routers are 3640s. Once I enabled logging for them I started receiving Bad Checksum messages showing up in the logs. This behavior is not consistant and not isolated to a few remote sites. The code and platform differs across the board. The code is pretty much 12.1.x across the board though. I show no physical layer errors on any of the interfaces nor do I see any buffer issues. The CPU and utilization of the routers both remote and host are at a minimum to say the most. On somedays I get 3 of these error messages and other days I may get 10 but never from the same host on a given day. In addition the messages are sometimes 15 mins. apart other times hours apart. OSPF topology is composed of 4 areas(remote sites) plus area 0(LAN backbone). If anyone has any ideas or maybe have seen this before your input would be appreciated.

[ruwhite]

First, try to narrow down which area this is happening in. Second, try to narrow down whether there is one adjacent router which is munging these packets, the local router is munging these packets, or the originator is munging these packets. To do this, look for patterns. If the munged packets are always received over the same interface, then you probably have a peer that is munging packets. If the munged packets are always from the same router id, then check that particular router out. Otherwise, it's probably the local router.

You have to remember that ospf routers do not just retransmit what they have received--they actually repackage the lsa's for transmission. So the packet could be corrputed by any router in the network.

:-)

Russ