We have recently upgraded our Internet connectivity from our data centre to include a 3rd peer with full Internet routing tables. Since then we have experienced a number of issues with routing loops on our core network. Inspection of the problem shows that everything should work correctly however traffic goes into a loop between two IBGP peers. On closer inspection we noticed the error:
%MLSCEF-SP-7-FIB_EXCEPTION: FIB TCAM exception for IPv4 unicast, Some routes will be software switched.
This indicated that the FIB table was not big enough on our MSFC2's in order to reduce the load we have implemented some BGP prefix filters on our accepted routes and reduced the amount of memory, however we are still experiencing problems.
Me questions are relatively simple:
1) Do we need to upgrade to SUB720-3BXL in order to cope with the increased route table size;
2) If we dont need to upgrade are there any further tweaks we could apply to ensure that we dont encounter these issues.
Thanks in advance for your help.