Description

Unexpected RE SWO (routing-engine switchover) happened on PTX-EVO while running a test with huge pim join/prune activity. No core was generated.

Symptoms

Nov 13 10:41:29.605 PTX-router re0-plat-mra[23396]: %USER-3-CHASSISD_RE_DISK_WRITE_RATE_EXCEEDED: RE[0] Disk(/dev/sda) write rate exceeded, rate/threshold: 5.360927/4
Nov 13 10:41:29.606 PTX-router re0-plat-mra[23396]: %USER-5-EMF_EVO_ALARM_SET: Alarm set: RE color=yellow, class=CHASSIS, reason=Host 0 Active Disk Write Rate Exceeded
Nov 13 10:41:34.384 PTX-router re0-jinsightd[21622]: %USER-3-JRESIL_MAINPROC_ERRITEM: Central Resiliency Entity (jResil) is processing an error with uri: /re/0/plat-mra/0/ssd/0/platform/0/Active Disk Write Rate Exceeded
Nov 13 10:41:35.976 PTX-router re0-jinsightd[21622]: %USER-3-JRESIL_MAINPROC_ACTION: Central Resiliency Entity (jResil) executed action: ReSwitchover

Solution

This issue is fixed in the below JUNOS EVO releases:

21.4R3-S5-EVO
24.1R1-EVO
23.3R2-EVO
23.4R1-EVO
22.2R3-S3-EVO
 
 
Workaround:
  • Reduce amount of disk writes (disable extensive traceoptions for example)
  • Change thresholds:
set chassis routing-engine disk write-threshold-rate <>
set chassis routing-engine disk write-threshold-duration <>

Modification History

2024-09-05 remove confidential PR number and add its non-confidential content.
2023-11-24 : Article Created