The message is generated by the chassis process (chassisd), which controls hardware components on the routing platform. The Packet Forwarding Engine divides packets into smaller units called cells for more efficient processing. As the indicated F chip on the indicated Control Board (CB) processed data before sending it to the indicated Packet Forwarding Engine on the indicated Flexible Port Concentrator (FPC) for outgoing transmission, it dropped the indicated number of cells per second.
The CHASSISD_FASIC_OUTPUT_DROP message is logged once per second for each F chip that is dropping cells meant for transmission to a specific FPC.
When a CHASSISD_FASIC_OUTPUT_DROP event occurs, a message similar to the following is reported:
chassisd: %DAEMON-3-CHASSISD_FASIC_OUTPUT_DROP: Fchip (CB 1, ID 0): dropped 50 cells per second destined for Packet Forwarding Engine 1 on FPC 9 chassisd: CHASSISD_FASIC_OUTPUT_DROP: Fchip (CB 2, ID 0): dropped 52 cells per second destined for Packet Forwarding Engine 0 on FPC 4 fpc5 MQCHIP(1) FI Reorder cell timeout
Note: The F chip dropped cells occurred only on DPF, not on MPC chipset.
The cause of the cell drops may be either traffic congestion on the F chip or the hardware being in a state that is preventing it from responding properly.
Run the commands below and examine the output to help determine the cause of the CHASSISD_FASIC_OUTPUT_DROP message:
- Look for any related events that occurred at or just before the CHASSISD_FASIC_OUTPUT_DROP message:
show log messages show log chassisd
- Look for any impending alarms for the fabric planes:
show chassis alarms
- Verify the fabric drops from the indicated PFE and FPC:
show pfe statistics traffic
- Look for any errors on fabric planes from the perspective of FPC:
show chassis fabric fpcs
- Look for any errors on PFE from the perspective of Planes:
show chassis fabric planes
- Check for fabric stats and drops as seen from source PFE (towards destination PFE), and monitor if the cell timeout occurred frequently in log messages (such as happening once per second):
show class-of-service fabric statistics
Perform these steps:
1. Reduce the amount of traffic going to the target FPC component.
2. During a maintenance window (as it will impact transit traffic), try the following:
- Reseat the control board in its slot.
- Move or swap the part, if possible, to an alternate slot.
- Replace the component with a spare.
3. If the failure messages continue, open a case with Juniper technical support representative to investigate the issue further.