ESXi host transport node preparation in NSX 4.2.2 is starting after a delay of two minutes.

Products

Issue/Introduction

In NSX version 4.2.2, upon adding a new ESXi host to the transport node profile applied NSX cluster, as seen in UI System > Fabric > Hosts > Clusters, the NSX preparation starts exactly after 2 minutes.

• Single MP Node Environment:

  • When a single host is added to a cluster, preparation commences after a 2-minute initial delay.
  • If multiple hosts are added, they are processed sequentially, each with a 2-minute gap between them. Therefore, the Nth host's preparation will begin at approximately the 2N-minute mark.
    
    

• Three MP Node Environment:

  • Hosts are processed in batches. The first batch (typically two hosts) begins after 2 minutes, the next after 4 minutes, and so on (Nth batch at 2N minutes).
  • It's important to note that while the observed batch size was two (due to internal sharding utilizing two MPs), this can vary from one (worst-case, only one MP processing) to three (best-case, all MPs processing simultaneously).
  • Consequently, the actual time for a host to begin preparation is influenced by the number of hosts already in the processing queue ahead of it.

In NSX 4.1.X, the host NSX preparation starts right away as soon as host is added to the NSX cluster.

Environment

• VMware NSX 4.2.2

Cause

The issue is affecting NSX version 4.2.2. 

Resolution

This issue is resolved in versions 4.2.3 available at Broadcom Downloads. If you are having difficulty finding and downloading software, please review the KB Download Broadcom products and software.

Please open a Broadcom Support case. For more information, see Creating and managing Broadcom support cases.