Description

This article explains the meaning of the alarm, 'Host x SMPC CPLD Golden Image Active' on MX series devices and how to resolve it.

This alarm may be seen after upgrading the Routing Engine or during the upgrade. One of the routing engines may trigger this error.

Symptoms

Active alarm:

user@host> show system alarms
1 alarms currently active
Alarm time               Class  Description
2019-04-25 16:50:58 UTC  Minor  Host 0 SPMC CPLD Golden Image active

Syslog Messages:

Feb  6 00:33:06.956 2020   alarmd[23623]: %DAEMON-4: Alarm set: RE color=YELLOW, class=CHASSIS, reason=Host 1 SPMC CPLD Golden Image active
Feb  6 00:33:06.981 2020   craftd[18436]: %DAEMON-4: Minor alarm set, Host 1 SPMC CPLD Golden Image active


Verify Hardware details:

user@host> show chassis hardware
Jun 21 03:38:07
Hardware inventory:
Item             Version  Part number  Serial number     Description
Chassis                                BLANK             JNP10003 [MX10003]
Midplane         REV 01   750-066883   BLANK             MX10003 Midplane
Routing Engine 0          BUILTIN      BLANK             JNP10003-RE1
CB 0             REV 07   750-067071   BLANK             JNP10003-SPM


Verify the Firmware version on  affected RE x:

Example:

user@host> show system firmware

Part                     Type       Version
RE 0                     PRI BIOS   CBEP_P_VAL1_00.14.01                       
                         FPGA       265.0                                      
                         RE-FPGA    42.0                                       
                         RE-SSD1    SF-SBR12050                                
                         RE-SSD2    SF-SBR12050                                
                         i40e-NVM   6.01                                       
RE 1                     PRI BIOS   CBEP_P_VAL1_00.14.01                       
                         FPGA       265.0                                      
                         RE-FPGA    42.0                                       
                         RE-SSD1    SF-SBR12050                                
                         RE-SSD2    SF-SBR12050  


Login to Linux Client ( VM Host ):

user@host>start shell user root
% vhclient -s

Verify FPGA version on RE that is problematic:

root@RE:~# re_fpga -V
    FPGA version   :    0x2a   <-- Verify this to be same on both REs
    JSPEC version  :    0x12
    PCB version    :    0x02
    Board Name     :    MX 10K

On problematic RE (Here, it is RE 0):
root@RE0:/var/log# re_fpga -r 0x1104
RE FPGA: Read reg 0x1104 -> Data 0x4f


On RE with no alarm:
root@RE1:/var/log# re_fpga -r 0x1104
RE FPGA: Read reg 0x1104 -> Data 0x17

Solution

  1. Corruption in SMPC FPGA.
  2. The device did not boot with the correct firmware.
  3. Unsuccessful firmware upgrade.

To clear the alarm, perform FPGA Firmware Upgrade.

The firmware should be updated to the latest available version. Refer to the technical documentation on Installing and Upgrading Firmware

Preferably, both REs should be on the same firmware version.

When the SPMC FPGA user image is corrupted and the good golden image is available, the following is expected:

  1. A message indicating "Booted from Golden FPGA"
  2. A "Minor Alarm" is seen
  3. Software can detect if it is booted from Golden copy and enable FPGA upgrade.

In certain cases, the alarm may be triggered due to a mismatch in RE versions as well. Ensure both the REs are on the same version, then check if the alarm clears. In the middle of an upgrade on dual RE devices, when both REs are on different versions, this alarm could be seen. In that case, upgrade the device with the best practice procedure, then check once both the REs are on the same page.

You may also see a BIOS mismatch/secure boot disabled alarm in addition to the existing alarm. Refer to KB35975 - Minor 'VMHost RE x Secure Boot Disabled' alarm [juniper.net]

From the affected RE x, collect the following logs (applicable for NGRE ):

user@host> start shell user root

​Enter Linux Host:

:~#vhclient -s

Collect VM Host Logs:

root@host:~# tar -cf ~/host_varlog_REx.tar -C /var/log/*
tar: Removing leading `/' from member names
root@host:~# ls -l | grep varlog
-rw-r--r--. 1 root root 75509760 Jun 18 02:55 host_varlog_REx.tar

Copy the file to VM Routing Engine  /var/tmp .

-- MX systems --
root@host:~# scp host_varlog_REx.tar [email protected]:/var/tmp/
Password:
host_varlog_REx.tar               100%   73MB  36.5MB/s   00:02


The above-collected VM Host logs will help to determine the root cause for the issue and for JTAC analysis. In addition to the above-collected outputs, you might need to have the following (preferable) for JTAC analysis:

  • Console Logs during boot up to check Golden Image load

Refer to Net Generation Routing Engine- NGRE for more details.

Please contact Juniper Support if the alarm does not clear after performing the above-mentioned steps for detailed analysis.