SDDC Manager vSAN Stretched Cluster Expansion fails at "Prepare Inputs for SubCluster configuration"

Products

• VMware SDDC Manager
• VMware Cloud Foundation 4.x
• VMware Cloud Foundation 5.x
• VMware Cloud Foundation

Issue/Introduction

• Failure During vSAN Stretched Cluster Expansion: "UNABLE_TO_GET_INPUT_FOR_NSXT_CONFIGURATION" at stage: "Prepare Inputs for SubCluster configuration"
• Error in /var/log/vmware/vcf/domainmanager/domainmanager.log

  ERROR [vcf_dm,67913992fe23fee86734441b756cda80,8cc3] [c.v.e.s.o.model.error.ErrorFactory,dm-exec-9]  [Q2HPFP] UNABLE_TO_GET_INPUT_FOR_NSXT_CONFIGURATION Unable to get input for NSX configuration
  com.vmware.evo.sddc.orchestrator.exceptions.OrchTaskException: Unable to get input for NSX configuration
  
  Caused by: java.lang.IndexOutOfBoundsException: Index 0 out of bounds for length

Environment

VMware Cloud Foundation 4.x

VMware Cloud Foundation 5.x

Cause

• The failure occurs because there are no sub Transport Node Profiles (TNP) or subclusters defined in NSX-T for a stretched vSAN cluster.
• This issue arises from an upgrade scenario where steps related to setting up sub TNP and subclusters were overlooked during the transition from 4.x to 5.x.
• With VCF 5.X it is mandatory and expected for a TNP and Sub-TNP to be configured on a stretched cluster for the sake of workflow consistency.

Resolution

• Due to the complexity of this issue, it is recommended to log a ticket with Broadcom support before proceeding with these instructions.
• Please note the following workaround steps should only be applied under the following conditions:
  • All hosts both in AZ1 and AZ2 are using the same host switch.
  • Overlay and vlan transport zones are the same on all hosts both in AZ1 and AZ2.
  • Uplink profile is configured the same on all hosts both in AZ1 and AZ2.
  • No IP pools are used for IP to TEP assignment. DHCP is only used for IP to TEP assignment in both AZ1 and AZ2.
• If the cluster does not match the above conditions, then the NSX configuration will need to be reviewed before appropriate steps can be provided.

Workaround:

1. Login to NSX Manager - (e.g. vcf-nsxvip.example.com)
2. Update TNP Associated with Cluster
   1. Identify and update the Transport Node Profile (TNP) associated with the cluster (Example: "vcf-shared-stretched") with a Sub TNP for AZ2.
   2. Provide the following details for Sub TNP:
      • Name: Example: vcf-shared-stretched-az2-nwp
      • Uplink Profile: (same as in global TNP)
      • IP Assignment: (same as in global TNP)
3. Create Sub Cluster for AZ2 Transport Nodes:
   • Create a sub cluster with the following details:
     • Name: Example: vcf-shared-stretched-az2-nwp
     • Hosts: Identify and add transport nodes in AZ2.
4. Re-apply the TNP on the cluster, mapping it to the newly created Subcluster and Sub-TNP.
5. Retry Failed Workflow in SDDC Manager

Refer Sub-TNPs and Sub-clusters for Step 2 and 3

Additional Information

If this issue is identified before initiating a stretch operation, refer to KB article 372306 for guidance on proper configuration.