I85PREP
⌘ K
Nutanix NCP-MCI study materials
Nutanix NCP-MCI study resources

Nutanix NCP-MCI: Troubleshoot and Maintain Cluster Health

Nutanix NCP-MCI · Work through your exam, one topic at a time.

All guide sections

Troubleshoot and Maintain Cluster Health

Troubleshooting items test the toolbox and the safe order of operations: diagnose first, then act, then verify - without turning a degraded cluster into an outage.

The toolbox

  • NCC: health checks returning PASS/WARN/FAIL; run from Prism or the CVM CLI. Scheduled daily, but run it on demand after any change.
  • logbay: collects logs into a bundle for support. Pulse sends telemetry to Nutanix.
  • CLI: cluster status for cluster-wide state, allssh/hostssh to run a command across CVMs/hosts, ncli and acli for configuration.
  • LCM for firmware and AOS upgrades; Prism Central must be at the same or newer version than the clusters it manages.

Safe operations

  • Put a host into maintenance mode before hardware work so VMs migrate and data rebuilds.
  • Restart a CVM with cvm_shutdown -P now, then power the CVM VM back on from Prism or IPMI - do not power-cycle the host.
  • Stop and start a whole cluster in the documented order, never by pulling power.

Worked example

A CVM is unreachable and Prism shows a degraded cluster. From another CVM, cluster status confirms one node's CVM is down while the host is up. You shut that CVM down cleanly, power it back on, wait for the metadata ring to re-form, then run NCC and confirm the checks return to PASS before closing the incident.

Traps

  • Running NCC as a fix - it only reports.
  • Rebooting hosts in production without maintenance mode.
  • Collecting logs after rebooting, destroying the evidence.
  • Upgrading Prism Element before Prism Central.