Cluster Health, Support, and Troubleshooting
Health checking
Nutanix Cluster Check (NCC) runs a library of read-only health checks across the cluster and reports pass/fail with guidance. It does not remediate anything. Runs are typically scheduled weekly and can be triggered manually from Prism or the CLI. Pulse is different: it phones home to Nutanix support with alerts and health data so issues can be triaged proactively.
Support tooling
When something is wrong, the flow is: review Prism alerts and the health page, run NCC, inspect tasks, then collect a support bundle and open a case. Logs can be gathered per node, and logbay collects them into a bundle. Day-to-day tools include ncli and acli for configuration and VM operations, allssh to run a command across CVMs, and hostssh to reach hypervisor hosts.
Troubleshooting patterns
Start with scope: one VM, one host, or the whole cluster? A single VM points at the guest, its networking, or NGT. A single host points at its CVM, disks, or hypervisor. Whole-cluster symptoms point at networking (VIP, bonds, VLANs) or consensus services. Data resiliency status tells you whether the cluster can survive another failure - check it before any maintenance.
Traps
Questions often credit NCC with fixing problems, or mix Pulse up with NCC. Beware of restarting cluster services as a first step; establish state and scope first.
Rule of thumb
Alert, then NCC, then support bundle, then case. Diagnose before you remediate.