Upgrades, Expansion, and Cluster Maintenance
This section covers Upgrades, Expansion, and Cluster Maintenance: upgrade planning and compatibility, adding and removing nodes, maintenance mode, rebalancing, and safe cluster changes.
Change rules
- Check compatibility and upgrade paths before moving AOS, AHV, firmware, or Prism Central versions.
- Run NCC and confirm free capacity before upgrades or node addition.
- Maintenance mode evacuates VMs from a host so it can be serviced; it is not the same as powering off a node.
- Adding a node requires compatible hardware, correct network addressing, Foundation imaging, and cluster expansion. Data rebalancing follows.
- Removing a node requires enough capacity and resiliency to absorb its data; verify RF and cluster state first.
Worked example
A four-node RF2 cluster needs a fifth node. The administrator confirms the new node is supported, assigns unique CVM and hypervisor IPs, images it with Foundation, and expands the cluster. After expansion, rebalancing distributes data. If the administrator instead removes a node first to free a slot, capacity and resiliency may drop; the safe order is to add capacity, verify health, then retire old hardware.
Traps that cost marks
Do not assume any node model can join any cluster. Do not skip NCC before an upgrade. Do not use maintenance mode as a substitute for proper shutdown. Do not remove a node without checking RF, capacity, and data placement.