NSX Edge Dataplane Fails after Upgrade or Upgrade of Hardware Version
search cancel

NSX Edge Dataplane Fails after Upgrade or Upgrade of Hardware Version

book

Article ID: 322410

calendar_today

Updated On:

Products

VMware NSX Networking

Issue/Introduction

Symptoms:
  • You have edge nodes deployed in Extra Large form factor.
  • You updated the edge hardware version to 17 or higher and are running NSX-T version 3.2 or lower.
  • The status of the edge is in UNKNOWN state System - Fabric -Nodes.
  • Checking /var/log/core/ on the edge node, there is a core dump for the dataplane service:
/var/log/core/core.datapathd.*.gz
  • In the edge log /var/log/syslog the below FATAL error is displayed:
FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="intel-rte" level="FATAL"] Requested more RX queues than max available for port 0max: 8 req: 9


Environment

VMware NSX-T

Cause

The Edge VM hardware version has been updated to version 17 (or higher).
In versions older than 3.2, control priority reserved a core for processing, and allocated 8 RX and TX queues for the VM.
In 3.2, an additional queue is used for control prio, and for extra large form factor edges there are not enough queues allocated by the ESXi for this request.
If the version is 3.2 and the edge form factor is Extra large, it is ok to upgrade the hardware version to 17 or higher and make use of the additional queue for control prio.

Resolution

This issue is resolved in VMware NSX 4.0, available at VMware downloads.

Workaround:
The workaround is to disable control prio on the edge node:
 > set dataplane ctrl-prio disabled
Please restart dataplane service for the change to take effect.
 > restart service dataplane
 
Note: Restarting the dataplane service on an edge node will impact traffic traversing the edge node, while the service is down.