Cluster Deployment, Installation, and Configuration
This section covers Cluster Deployment, Installation, and Configuration: Foundation imaging, IP planning, CVM and hypervisor addressing, cluster creation, initial Prism configuration, and the checks that make a cluster supportable.
Deployment sequence and rules
- Foundation discovers nodes, often via IPMI or MAC, and images the hypervisor and AOS.
- Plan unique CVM IPs, hypervisor management IPs, a cluster virtual IP, subnet mask, gateway, DNS, and NTP before imaging.
- CVMs must reach each other on the cluster network; management access must reach Prism and hypervisor hosts.
- Configure NTP and DNS early. Time skew and name resolution faults cause authentication, alerting, and upgrade failures.
- After cluster creation, run NCC health checks and confirm licensing, alerts, and Prism access.
Worked example
You image three nodes for an AHV cluster. Each node needs a unique CVM IP and a unique AHV management IP; the cluster VIP must be unused. All addresses must be in the correct VLAN and subnet, with a valid gateway. If two nodes receive the same CVM IP, discovery may appear to work but cluster formation fails or is unstable. The fix is to correct the address plan in Foundation before cluster creation, not to patch CVM files afterward.
Traps that cost marks
Do not reuse the cluster VIP as a CVM or hypervisor IP. Do not skip NTP because the cluster looks healthy at first. Do not assume Foundation configures every VLAN or bond; those may need later Prism or network setup. Do not treat NCC warnings as optional during handover.