Showing results for 
Search instead for 
Did you mean: 

Core issue

In this case, Cisco Express Forwarding (CEF) is disabled due to a low memory condition. If the ip cef [distributed] command is configured, CEF can be re-enabled.

Cisco recommends the reduction of other system activity in order to ease memory demands. If conditions warrant, upgrade the Route Processor (RP) so that it contains more DRAM.

In some scenarios, flushes on the input queue of an interface relates to full buffers.

This can be explained as a packet arrives on the interface and is received by the local controller, the interface driver receives the packet and checks for basic data integrity, if buffers are available the packet is processed normally. If there are insufficient local buffers available, the interface driver tries to allocate space from the public particle pools in DRAM/SDRAM. If there are no free buffers available in the public pool, the interface drive flushes the frame and increments the ignore and no buffer counters.

If the router is unable to forward the packet with fast/CEF switching, the buffer particles are coalesced and the packet is placed into a system buffer for process - switching. The packet is now enqueueed and the interface hold queue is incremented. If the hold-queue is full, the packet is discarded, and the input queue drop error is generated.

The %FIB-2-FIBDOWN: error message can occur due to these reasons: 

Reason 1:  When the number of not-directly connected routes that the desktop default template allows is exceeded, this error message can be received. If this template is used, it is likely that the maximum number of 2000 is exceeded.

Reason 2:  When the number of MAC addresses learned by the switch has exceeded the amount of space allocated in hardware to store MAC addresses, this error message can be received. In this case, the show mac-address-table count output shows 0 free entries.

Reason 3:  If the Catalyst 3750 experiences a memory leak in the IP Routing Information Base (RIB) update process, this error message can be received. The process begins to hold more and more memory, and does not release it.


Try these workarounds based on the reasons stated in the Core Issue section:

Workaround for Reason 1:  Issue the sdm prefer routing command and reload. Ideally, this resolves the problem.

Refer toConfiguring SDM Templates for more information.

Workaround for Reason 2:  Change the Switch Database Management (SDM) template in order to allow for more space in the unicast MAC address region, as instructed in Cisco bug ID CSCef89559.

Workaround for Reason 3: Reboot the switch as instructed in Cisco bug ID CSCef35924.

Refer to article User is unable to configure PBR on a Catalyst 3550 switch for more information.

In case of 6000/6500 switch, if the reboot does not help, issue the ip cef linecard ipc memory 10000 global command in order to configure the line card memory pool for the CEF queuing messages.

Error message


Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Recognize Your Peers
Quick Links