AHV - Shutting down the cluster​ - Maintenance or Re-location

  • 15 November 2019
  • 0 replies
  • 4723 views

Userlevel 3
Badge +4

Nutanix offers a distributed data and control plane, so it’s fairly easy to start, stop and graceful shutdown a cluster. Even, in abnormal / dirty shutdowns, Nutanix cluster has powerful self-healing capabilities - as all data & meta-data is distributed across the cluster which significantly reduces the chances for data corruption or data-loss. 

However, as a Nutanix Cluster hosts business critical data and applications, it is important to ensure all services stop in a graceful manner and all data + meta-data is consistent. This allows the cluster to be restarted in a healthy - usable state later.

There can be several reasons to gracefully shutdown a running Nutanix AHV & AOS Cluster. When shutting down a Nutanix Cluster, following order needs to be followed:

  1. User VMs Shutdown
  2. AOS Shutdown - Data services / Cluster components
  3. CVM Shutdown
  4. Hypervisor Shutdown

In this post, we will focus on the shutdown process for a Nutanix Cluster running with AHV (Acropolis Hypervisor).

Points to Check and Consider:

  1. Check for any running replications on your Nutanix Cluster > Data Protection
  2. Check for any running backups and ensure they finish successfully
  3. There are no running software or firmware upgrades in progress for AOS
  4. Run ncc (nutanix cluster check) utility and ensure there are no red flags / alerts
  5. Do you use Nutanix Files? then Files (AFS) will need to be shutdown first.

 

Steps to Shutdown - AHV & AOS Cluster


Prior to the scheduled shutdown, SSH to a Controller VM and run:
ncc health_checks run_all. If there are any errors or failures, Check the relevant KBs and contact Nutanix Support if required.

  1. Shut down all the user VMs in the Nutanix cluster
  2. Stop all AFS cluster VMs if applicable
  3. Stop the Nutanix cluster.
  4. Shut down each node in the cluster.
  5. After completing maintenance or other tasks, power on the nodes and start the cluster.

Please visit the “Shutting Down an AHV Cluster for Maintenance or Relocation” KB to read more about how you can shutdown several user virtual machines with a single command.


This topic has been closed for comments