FEB * FPC link errors
This alarm is generally raised when there are HSL link errors between the FEB fabric planes and FPCs.
The HSL link is the backplane connection between the FPC and FEB. When the link does not function as expected, the issue is reported by the H2S ASIC located at the FPC. This ASIC is responsible for communication between I-chips located at the FEB and FPC.
The problem also could be due to a connection issue between the FEB and the FPC, or either hardware component might be in a state that is preventing the connection from performing as desired.
Perform these steps to determine the cause and resolve the problem (if any).
Collect the show command output.
1) Capture the output to a file (in case you have to open a technical support case). To do this, configure each SSH client/terminal emulator to log your session.
show log messages | no-more
show log chassisd | no-more
show chassis fabric map
show chassis fabric feb
show chassis fabric plane
show chassis fabric plane-location
show chassis fabric summary
show chassis environment feb
show chassis fpc-feb-connectivity
show pfe statistics error feb [#]
Be sure to replace the [#] above with the number of the FEB associated with the system message. You might need to run the show pfe statistics error feb [#] command multiple times if more than one FEB is reporting an issue, as seen in the show log messages output. The fabric outputs help to determine which fabric connection or connections are reporting errors.
2) Analyze the show command output.
Check the output from show log messages to see if just one FEB is reporting the issue, or if multiple FEB components are reporting, and also if just one or more FPC units are showing the link errors. This will help narrow down which part or parts to concentrate on. The show chassis alarms output might provide an additional clue as to where to look.
3) Check the following outputs from the FPC and FEB boards. These are also useful in investigating this issue. Examine the logs from each FPC/FEB referred to in the system message with one of the following sets of commands:
start shell pfe network [fpc#/feb#]
#show syslog messages
#show nvram
/* M120 FPC Type-2,3 */
#show hsl2 asic h2s0
#show hsl2 asic h2s0 serdes
/* M120 FPC Type-1*/
#show hsl2 asic bdh0 serdes
#show hsl2 asic bdh0
#show hsl2 statistics
#show hsl2 statistics detail
#sh conv_fpga info
> exit
-OR--
request pfe execute command "show syslog messages" target [fpc#/feb#]
request pfe execute command "show nvram" target [fpc#/feb#]
request pfe execute command "show hsl2 statistics" target [fpc#/feb#]
......
Be sure to replace the [fpc#/feb#] entry above with the reference to the system component that is reporting the log message. These logs will show which components are logging a link error.
4) During a maintenance window, as these steps will affect transit traffic, try the following to resolve the link error:
Look for one FEB or FPC to be in common in the system messages, and try resetting that component. This can be done by taking the component offline briefly, and placing it back online. (If there is not one component in common, then this will need to be done on all the components referenced in the system messages.)
If resetting the one component resolves the issue temporarily, but it returns a few hours later, then open case with your technical support representative to investigate the issue further.
If a reset does not resolve the issue, try reseating the component in its slot.
Swap the FPC and FEB components with another, identical component in the chassis, one at a time, to see if the messages follow the FPC that moved or the FEB that is moved.
It is possible to change the association of a FEB with an FPC in the configuration under chassis settings. Try swapping the FPC associations with the FEB components to see if the messages follow the FPC or the FEB. Example configuration commands to use could be:
set chassis fpc-feb-connectivity fpc 1 feb 0
set chassis fpc-feb-connectivity fpc 0 feb 1
If you have spare FEB and/or FPC components available, try swapping them to see if replacing one component resolves the issue.