PXE boot platform VMs and expand to 3-node HA

[jmgilman]

Boot the platform VMs via PXE to install Talos and join the platform cluster.

Prerequisites:

• CP-2 and CP-3 VMs created on Harvester
• Hardware XRs registered in Tinkerbell
• Talos machine configs available (cp-2.yaml, cp-3.yaml)

Tasks:

• Power on CP-2 VM in Harvester

• Verify PXE boot and Tinkerbell workflow execution

• Wait for Talos install and cluster join

• Power on CP-3 VM

• Verify PXE boot and cluster join

• Verify platform cluster has 3 control plane nodes:

  kubectl get nodes
  # Expected: platform-cp-1 (UM760), platform-cp-2, platform-cp-3
  

• Verify etcd quorum (2/3 required)

• Test HA by cordoning one node and verifying cluster remains operational

Expected Duration: ~20 minutes (VMs can boot in parallel)

Result: Platform cluster is now highly available (3-node etcd quorum)

Reference:

• docs/architecture/appendices/B_bootstrap_procedure.md (Step 17)

[linear]

HOM-15