Monitoring NSX Edge disk usage in VMware NSX for vSphere 6.1.x and higher

Products

• VMware NSX Networking

Issue/Introduction

This document describes the key partitions of the NSX Edge disk and provides suggested disk usage thresholds as configurable inputs for monitoring applications in VMware NSX for vSphere 6.1.x and higher.

Environment

VMware NSX for vSphere 6.2.x
VMware NSX for vSphere 6.3.x
VMware NSX for vSphere 6.4.x
VMware NSX for vSphere 6.1.x

Resolution

In VMware NSX for vSphere releases 6.1.1 and higher, the Edge virtual machine contains four partitions:

• root - the root partition does not store the configuration database or other run-time files, such as PID files, control sockets, and DHCP lease files. Therefore, utilization is expected to remain consistent.
  
  Note: VMware recommends an alert threshold of 95%.

• db – Though the db (/var/db) and log partitions are relatively small, they store dynamic data and should be monitored more closely than root to avoid an out-of-space condition.
  
  Note: VMware recommends a warning alert threshold of 90% and a critical alert of 95%. Setting lower thresholds limit the usable space on an already small partition.

• dumpfiles - The dumpfiles partition is used only to store tcpdump files and core files if core dumps are enabled.
  
  Note: By default, the dumpfiles partition is small because enabling core dumps creates a second vmdk mounted at /var/dumpfiles to store core files. You may choose to monitor this partition when tcpdump files are saved during debugging sessions.

• log – VMware recommends a warning alert threshold of 90% and a critical alert of 95%. Though the Edge continues to operate, filling this partition effectively prevents new logs from being saved to disk.