VxRail hosts are not discovered and cannot be added when IPv6 is disabled on the hosts
search cancel

VxRail hosts are not discovered and cannot be added when IPv6 is disabled on the hosts

book

Article ID: 448710

calendar_today

Updated On:

Products

VMware SDDC Manager / VCF Installer

Issue/Introduction

  • You attempt to add or migrate VxRail hosts into an existing workload domain cluster, and no hosts are discovered.
  • The SDDC Manager Add Host wizard fails at the Discovered Hosts step with an error similar to:
    Failed to load Host cluster details. Http failure response for
    https://<sddc-manager-fqdn>/ui/api/v1/clusters/<cluster-id>/hosts/queries/<query-id>: 500 Internal Server Error
    
  • In the vCenter VxRail plugin, under Cluster > Configure > VxRail > Hosts, you see:
    The VxRail Manager system information is not found.
    
  • The Add VxRail Hosts wizard shows an empty discovered-hosts table with:
    0 hosts discovered.
    
  • When you try to add a host manually by IP address, validation fails with the host flagged Not Compatible and blank Service Tag, PSNT, and Model columns:
    Cannot find node with provided IP address and password
    

Environment

  • VMware Cloud Foundation 5.2.x on Dell VxRail
  • ESXi 8.0
  • VxRail 8.0

Cause

VxRail node discovery is handled by the loudmouth service, which relies on IPv6 to detect nodes. The service uses each host's IPv6 link-local address (fe80) and IPv6 multicast on the VxRail private management network so that VxRail Manager can find the nodes automatically.

When IPv6 is disabled on the hosts, they no longer present the IPv6 link-local address that discovery depends on, so VxRail Manager cannot detect them. SDDC Manager builds its host inventory by querying VxRail Manager, so the absent node data causes the discovery query to fail and return the 500 Internal Server Error, while the VxRail plugin reports that VxRail Manager system information cannot be found and that zero hosts are discovered.

The Not Compatible flag and the blank Service Tag, PSNT, and Model columns are downstream effects of the nodes being unreachable to the wizard, not an actual software or hardware version mismatch.

Resolution

  1. In the vSphere Client, select a host you are trying to add and go to Configure > Networking > TCP/IP configuration. Confirm whether IPv6 is enabled on the host.
  2. If IPv6 is disabled, enable IPv6 support on the host. This setting requires a reboot to take effect, so evacuate or safely power down workloads as needed and reboot the host.
  3. After the reboot, confirm the management VMkernel adapter (vmk0) has an IPv6 link-local address beginning with fe80.
  4. Restart the loudmouth node-discovery service so discovery runs again cleanly. For the exact service commands, see the Dell reference in Additional Information.
  5. In vCenter, reopen the VxRail Add VxRail Hosts wizard and confirm the nodes now appear in the discovered-hosts list.
  6. Continue with the cluster expansion to add the nodes.

If the hosts are still not discovered after IPv6 is enabled:

  1. Confirm the VxRail private management network carries IPv6 multicast between the nodes and VxRail Manager, and that IPv6 and multicast are not being filtered on the top-of-rack switch ports for these hosts. Nodes reset with RASR revert to the factory discovery VLAN, which the switch ports may no longer carry after the nodes were in production.
  2. Confirm the management VMkernel adapter is not holding an unwanted auto-configured IPv6 address (fc00: prefix). If it is, follow the related Dell article in Additional Information to correct the IPv6 auto-configuration setting.

Additional Information