Scaling out Prism Central

Scaling out adds capacity and resilience, at the cost of two more VMs and a decision that cannot be undone. The two high availability mechanisms below do different jobs, and both can apply.

Scale-out rules

Before starting scale-out, confirm the gateway is reachable, addresses are unique and the container is mounted on the hosts. On ESXi, vCenter must be correctly configured, registered in Prism, reachable, and have DRS enabled.

The procedure is Settings > Prism Central Management > Scale Out PC, add the virtual IP, enter the two addresses, then Expand. If a proxy is configured on the cluster, add the Prism Central and Prism Element addresses to the proxy allowlist. At a dark site, where the image cannot be reached, upload the 1-click deploy files to a CVM and run the command below, then continue.

nutanix@cvm$ ncli software upload \
  software-type=PRISM_CENTRAL_DEPLOY \
  file-path=file_path \
  meta-file-path=metadata_file_path

Three follow-up tasks apply after expansion:

• Set up VM-VM anti-affinity rules between the Prism Central VMs.
• Run an LCM inventory if a linked product, especially NCM Self-Service, shows a version mismatch.
• If application monitoring ran on the single-node setup, power the new nodes off and on.

Checking resources before and after changes

The Prism Central Allocation and Consumption tab, under Prism Central Management, is the single place to review the instance. It shows total vCPU and memory consumption over the last hour, 24 hours or week, vDisk capacity per VM, and how vCPU and memory were allocated originally and to each service. Check it before enabling a feature.

If VM resources are below the amounts in KB 8932, manually increase memory and vCPUs. Only supported additions belong on the VM: do not add vDisks or NICs beyond those the installation or upgrade instructions specify.

Deleting any scaled-out Prism Central VM can result in data loss, so treat them as one instance. The console also supports a maximum of 10 concurrent user sessions across UI and API combined.

Common trap: Spreading the three VMs across two clusters to survive a cluster failure - all scaled-out VMs must run on the same cluster, so scale-out protects against losing one VM, not a site. See “Protecting Prism Central” for cluster-level recovery.

Get the whole book

This note is one section of Ultra Transcenders NCP-MCI-7.5: Nutanix Multicloud Infrastructure, an independent study guide that explains every topic the exam covers by technology, with comparison tables, diagrams and the common traps, plus a glossary linked to the Nutanix Support Portal.

About the book · NCP-MCI-7.5 terms in the glossary · All NCP-MCI-7.5 study notes