Resize NSX Edge transport Node

Products

• VMware NSX
• VMware NSX-T Data Center

Issue/Introduction

• This document explains how to resize an NSX Edge node with a higher form factor.
• This is applicable when Support has advised to resize an NSX Edge node due to performance issues, resource usage exceeding the threshold or any other reason.
• Note:
  • This present article is applicable only in non-VCF deployed NSX environments, using the NSX Management UI.
  • For VCF, Refer to the following article for Edge node resize tool: VMware Cloud Foundation NSX-T Edge Node Virtual Machine Resizing Tool.
  • Alternatively, NSX Edge nodes can be replaced/resized via API call: Using API calls to Redeploy Edge Node.

Environment

VMware NSX

Cause

The need to resize an NSX Edge transport node to a higher form factor is typically triggered by one of the following scenarios:

• Performance Issues: The current Edge node is underperforming and struggling to handle the network workload.
• Resource Exhaustion: The Edge node's resource usage (CPU, memory, etc.) is consistently exceeding established thresholds.

Resolution

Note: Edge node deployment and redeployment will fail due to a known issue, apply the workaround in the following KB before proceeding.  "OVF certificate validation failed. Error: [VALIDATION_ERROR: CERTIFICATE_EXPIRED; ]" error for NSX Edge Install/Redeploy/Resize.

For environments running VMware NSX 4.2.2 or later, resize Edge Transport Nodes directly through the NSX UI. This method seamlessly retains the existing Edge node FQDN and IP addresses. For step-by-step instructions, Refer the Redeploy an NSX Edge Node section of the NSX Administration Guide:

1. To minimize environmental impact, this operation preferred to be executed during a scheduled maintenance window. Administrators must ensure a valid NSX backup is available before initiating the Edge transport node resizing process.
   1. During the resizing workflow, the Edge Transport Node enters maintenance mode. If the node is part of an Edge cluster, its active connections will automatically fail over to other available nodes in the edge cluster with minimal disruption.
2. Redeployment workflow performs the following operations:
   1. Places the existing edge node in maintenance mode and then powers off the node.
   2. Deletes the existing edge node VM.
   3. Deploys a new edge node VM with the same configuration settings that specified.
3. Log in to the NSX UI using admin account. Navigate to System > Fabric > Nodes > Edge Transport Nodes.
4. Select the particular Edge transport node to resize to a higher form factor. Click on ACTIONS > Redeploy Edge. 

    3. Choose the required Form factor size as per the requirement and read the information popup and click on "NEXT".

   4. Enter the admin and root account credentials and click on Next.

    5.  No modifications are required within the "Configure Deployment" section, as the existing configuration is retained and displayed by default. Click on "NEXT".

    6.   No modifications are required within the "Configure Node Settings" section, as the existing configuration is retained and displayed by default. Click on "NEXT".

    7. No modifications are required within the "Configure NSX" section, as the existing configuration is retained and displayed by default. Click on "FINISH".

    8. This will take couple of minutes to finish the task.

    9. Repeat steps 1 through 8 for all subsequent Edge Transport Nodes. To ensure continuous environment stability, perform this operation on a single node at a time.

 

For NSX versions prior to 4.2.2, or if the direct UI resize method is unavailable, use the following procedure to replace the existing Edge nodes in the cluster with new node of a higher form factor:

• This is 3 Stage process.
  • Stage 1: Deploy new Edge Nodes in desired size.
  • Stage 2: Configure NSX.
  • Stage 3: Replace Edge Node within the edge cluster.

• Stage 1: Deploy new Edge Nodes in desired size:
  • Login to NSX UI as admin user and select System > Fabric > Nodes > Edge Transport Nodes and click "ADD EDGE VM".
  • This is an example to replace "Large" edge nodes with "Extra Large" , hence selected the new node as "Extra Large" in size:

• • Specify the passwords for admin, root and audit user accounts and Click "NEXT".
  • Select the Compute manager, Cluster, NSX resource pool and Datastore to which this edge will be deployed and Click "NEXT":

• • Select static IP address and specify the following configuration:
    • Management IP of the Edge VM and gateway IP, Management Interface, Search domain name and select the domain name, Enter DNS Servers details, Enter NTP/Time servers details. Click "NEXT":

 

• Stage 2: Configure NSX:
  • Note the following example is a specific to older environments that have been upgraded from NSX 2.x as it is configured with N-VDS rather than native VDS so two NSX Edge node switches are configured.
  • Configure the following settings:
    • Edge switch name, Transport Zone, Uplink Profile, IP Assignment, IP Pool, Virtual Nic.

• • Select Add switch and configure the following settings:
    • Edge switch name, Transport Zone, Uplink Profile, Virtual Nic, Click "Finish"

 

• Stage 3: Replace Edge Node within the edge cluster:
  • Select System > Nodes > Edge Clusters. Select the "Edge cluster". Click actions and "Replace Edge Cluster Member":

• • Select the current edge node from "Replace"  and then select the new Extra Large Edge from "With" section to replace edge node:

• The above process will replace the Edge Node in the cluster and also update the T0 interface to associate the new Edge with the uplink IP address used with the previous Edge. Below is the procedure to confirm the T0 is updated.
• This can be validated as follows:
  • On NSX UI, Select Networking > Tier-0 Gateways.
  • Expand the T0 gateway > Expand Interfaces and select the number next to external and service interfaces to display the uplink interfaces:

• We have swapped the old Edge with a new Edge. This can take several minutes so keep refreshing until the new edge is displayed. If there is a DNS record associated with this edge node, Update the record to reflect the new FQDN:

• Repeat the Edge node replacement on the other nodes one by one.
  • Once all nodes have been updated and we can see that the smaller nodes are no longer part of the cluster, we should delete the smaller nodes.

Additional Information

Note: Do not change the Compute Manager setting. Modifying this setting permanently erases all configuration mappings associated with the current compute manager. These mappings cannot be recovered, even if revert to the original compute manager.

If additional assistance is needed with troubleshooting, reach out to Broadcom support by creating a support case using the instructions at Creating and managing Broadcom support cases.