Introduction
This document describes how to fix SMAUG – Field Programmable Device (FPD) "NOT READY" and the logs that need to be collected for Root Cause Analysis (RCA).
Troubleshoot "SMAUG FPD NOT READY" in NCS4K
Platform: NCS4K
Card: 400G Line card – NCS4K-4H-OPW-QC2
Probable Trigger: FPD upgrade and card reload, RP reload, or switch over
The issue is depicted here:
Problematic state is depicted
Logs to be collected:
- From XR:
show tech-support stat
andshow tech-support slice-manager
- From Admin:
show tech-support slice-manager
, show tech-support shelf_mgr
,show tech-support ccc
, show tech-support HBloss
- From the XR:
dumpcore running grima_driver location <active LC>
- From the active LC get the traces for grima and dma. Perform the steps from XR Prompt.
#attach location 0/lc0
Fri Jun 7 00:20:00.329 CDT
[xr-vm_node0_LC0_CPU0:~]$ export PS1=#
#lspci -vd:8bec
#grima_ltrace_show -A
#nl_dma_ltrace_show -A
To recover the issue:
- From XR check for the process for
grima_driver
and slice_manager
, then restart them for active LC location.
show processes grima_driver location <0/LC0 or 0/LC1 – active LC>
show processes slice_manager location <0/LC0 or 0/LC1 – active LC>
Process restart grima_driver location <active LC>
Process restart slice_manger location <active LC>
This must fix the issue in most cases.