In my case I had a lot of CRC errors on the server NIC. Changing the NIC speed to match the setting on the Cisco switch solved my problem.
This problem occurs because of a remote procedure call (RPC) issue. An RPC server may incorrectly cancel all later asynchronous RPC calls after an RPC client cancels an asynchronous RPC call. See ME930850
for a hotfix applicable to Microsoft Windows Server 2003.
This event can have multiple causes. See ME911799
for an article that may resolve your problem.
The routing table must be adjusted correctly. In my case, after removing unnecessary records, the problem was solved.
This event can also appear when promoting a new DC, in a site with a small bandwidth. Running dcpromo /adv and choosing restore from backup could stop this event from occuring.
As pee Microsoft: "This issue may occur if you are trying to replicate Active Directory changes that cannot be replicated in five minutes or less. To resolve this issue, increase the bandwidth of your network connection so that the Active Directory changes replicate in the five minute timeout period. If you cannot increase the bandwidth of your network connection, edit the registry on your Windows Server 2003-based computer to increase the value of the RPC timeout for Active Directory replication". See ME830746
for more details.