FRR log file size increases to a large size - causing Edge instability from high resource usage

Products

• VMware NSX

Issue/Introduction

Symptoms:

• NSX Edge VM CPU usage is occasionally high or very high (seen in the alarms below):
  • Edge CPU usage high alarm
  • Edge CPU usage very high alarm
• Below errors are observed very frequently in /var/log/syslog:
  • YYYY-MM-DDThh:mm:fffZ <edge_name> rsyslogd - - -  file '/var/log/frr/frr.log': open error: Permission denied [v8.2304.0 try https://www.rsyslog.com/e/2433 ]

Environment

VMware NSX versions 4.1.0 through 4.2.1.2 (inclusive)

Cause

The log rotation of frr.log is not functioning correctly in the impacted versions.
Memory consumption by the Edge node grows, eventually leading to an unstable Edge (Edge running out of space).
An unstable Edge can lead to potential traffic/workload disruption.

Resolution

This issue is resolved in NSX version 4.2.1.3.

VMware NSX 4.2.1.3 Release Notes : Fixed Issue 3480984