Syslog error XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR (0x230062)

On all MX platforms having MPC10 or MPC11 having subscriber services with class-of-service configured, it is observed that in a scaled scenario (1500 IFLs (Interface Logical)), when queues are oversubscribed and the output interface starts to get congested, "CMERROR 0x230062" or "XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR" error message is seen. These cm errors cause the PFE (Packet Forwarding Engine) to be disabled.

Here is a snippet of logs for reference:

Jul 17 20:56:37 mx960 : %PFE-5: fpc0 CMError: /fpc/0/pfe/0/cm/0/XQSS(1)/1/XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR (0x230062), scope: pfe, category: functional, severity: major, module: XQSS(1), type: XQSS sched QL4 DQU QSUM overflow, oc_category: unknown oc_category
Jul 17 20:56:37 mx960 : %PFE-5: fpc0 Performing action log for error /fpc/0/pfe/0/cm/0/XQSS(1)/1/XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR (0x230062) in module: XQSS(1) with scope: pfe category: functional level: major, oc_category: unknown oc_category
Jul 17 20:56:37 mx960 : %PFE-5: fpc0 Performing action get-state for error /fpc/0/pfe/0/cm/0/XQSS(1)/1/XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR (0x230062) in module: XQSS(1) with scope: pfe category: functional level: major, oc_category: unknown oc_category
Jul 17 20:56:37 mx960 inetd[13651]: %DAEMON-4: Number of tftp connections at max limit (1)
Jul 17 20:56:37 mx960 chassisd[13643]: %DAEMON-3-CHASSISD_FPC_ASIC_ERROR: <FPC 0> ASIC Error detected errorno 0x00230062 (null)
Jul 17 20:57:04 mx960 : %PFE-5: fpc0 Performing action cmalarm for error /fpc/0/pfe/0/cm/0/XQSS(1)/1/XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR (0x230062) in module: XQSS(1) with scope: pfe category: functional level: major, oc_category: unknown oc_category
Jul 17 20:57:04 mx960 : %PFE-3: fpc0 Cmerror Op Set: XQSS(1): XQ-SS[1]: SCHED DQU QSUM overrrun interrupt (URI: /fpc/0/pfe/0/cm/0/XQSS(1)/1/XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR)
Jul 17 20:57:05 mx960 alarmd[14852]: %DAEMON-4: Alarm set: FPC id=150995048, color=RED, class=CHASSIS, reason=FPC 0 Major Errors
Jul 17 20:57:05 mx960 craftd[13646]: %DAEMON-4: Major alarm set, FPC 0 Major ErrorsJul 17 20:57:17 mx960 : %PFE-5: fpc0 user.notice root: pfe_get_state.sh : Executing CLI commands from /var/tmp/.__pfe_debug_commands
Jul 17 20:57:17 mx960 : %PFE-5: fpc0 user.notice root: pfe_get_state.sh : PFE get state execution successfully completed - /var/tmp/.__pfe_aft_debug_info_SLOT0.1

FMPC0(mx960 vty)# sh cmerror module 18         
Module (18) (XQSS(1))
PFE support 1
Get state cb 0x0 platform action cb 0x8e3bef0 control cb 0x0
Error-id  PFE Level Threshold Count Occured Cleared Last-occurred(ms ago) Name
------------------------------------------------------------------------------------------
0x230062  1 Major  1   0   1   1   21069125     XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR

FMPC0(mx960 vty)# ... module 18 error 0x230062 pfe 1  
Error-id      : 0x230062
Error Name       : XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR
Identifier       : /fpc/0/pfe/0/cm/0/XQSS(1)/1/XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR
Description       : XQSS sched QL4 DQU QSUM overflow
State          : enabled
Scope          : pfe
Category        : functional
PFE           : 1
Configured Level    : Major
Default Level      : Major
Count          : 0
Threshold        : 1
Error Limit       : 0
Occur Count       : 1
Clear Count       : 1
OverItemThres Occur Count : 1
OverItemThres Clear Count : 1
Last-occurred(ms ago) : 21168170

Logs:
----------------------------------------------------------
Index Time        Sub-Err State Description
----------------------------------------------------------
0  07/18/24 07:53:14 0    Clear unknown --> after clearing fpc errors during non-impacting scenario
1  07/18/24 02:57:04 0    Set  unknown

Under scaled conditions i.e. with 1500 IFLs on one side sending and receiving traffic with all queues on these IFLs oversubscribed and class-of-service configured, we are seeing cm alarm "CMERROR 0x230062" or "XQSS_CMERROR_SCHED_QL4_INT_REG_DQU_QSUM_OVR". As a result, of these errors the PFE remains disabled.

This issue might be seen if the following conditions are met:

* On all MX platforms with MPC10/MPC11 having subscriber services

* Class-of-service configured

* Scaled scenario with 1500 IFLs on one side sending traffic and receiving traffic with all queues on these IFLs oversubscribed

The issue is fixed via PR1793375

https://prsearch.juniper.net/problemreport/PR1793375