If you want to shutdown Nutanix AHV cluster for periodic maintenance or hardware relocation or any down time etc. You need to follow the predefined Nutanix cluster shutdown procedure.
This is very rare activity to shutdown or power off the Nutanix Acropolis AHV cluster in production environment.
Lets do this practically using Nutanix Prism Web console and Command line interface (CLI)
Step 1: Log in to Nutanix Prism in web browser
Note : Check Nutanix AHV cluster “Data Resiliency” status it must be “OK” if NOT, please wait until it turns “OK” Nutanix has auto healing system.
If does not turn “OK”, Please Log a case with nutanix or hardware vendor support.
Step 2: Precheck the Nutanix cluster health from Prism web console
Option 2.1. Go to the Menu > Health > Action> Run NCC checks > All Checks (*) > Run
Optional 2.2 you can also run Nutanix cluster heal check from Command line interface (CLI)
Through SSH (Using Putty/Linux terminal) to any CVM (Controller VM) of Nutanix AHV cluster and run following command and Wait until run successfully
cvm$ ncc health_checks run_all
If something is alarming critical alert then Contact to nutanix or hardware vendor support if everything is fine go ahead to step 3.
Step 3 : Shutdown all guest VMs except Controller VM – CVM
To shutdown Nutnanix Guest VMs has two options
Option 1 : From Nutanix Prism Web console
Option 2 From acli – command line interface (CLI)
Just login to any CVM (Controller VM) through SSH of Nutanix AHV cluster and run following command
CVM$ acli vm.shutdown * |For soft shutdown| OR cvm$ acli vm.off * |For hard shutdown|
Note : “*” for all guest VMs will shutdown except CVM
Now exit from acli command interface and proceed to run next command
Step 4: Stop / shutdown Nutanix AHV cluster through following command
CVM$ cluster stop
Command will prompt confirmation type “YES” to proceed to shutdown cluster. It will take up to 1 minute.
Step 5: Now shutdown controller VM – CVM
Log in to each CVM (Controller VM) one by one of Nutanix AHV cluster and run following command:
cvm$ cvm_shutdown OR cvm$ cvm_shutdown -P now
Note : After all CVM goes shutdown go ahead for step 6
Step 6: Now shutdown AHV node, one by one through IPMI or SSH
There is two option to shutdown the Nutanix AHV host
Option 1 From IPMI web interface
Browse IPMI IP address in chrome / Firefox / Opera / Safari / IE etc.
Go to Remote control > Power control > Power Off Server - Orderly Shutdown
Option 2 From Command line
SSH to each AVH host one by one of Nutanix AHV cluster and run following command:
host# init 0 |Note: 0 (Numeric zero)|
Confirmation : To check AHV power on/off status just ping the AHV host IP address or on AHV host IPMI home page power status
Power on the Nutanix AHV nodes and cluster after a shutdown
Step 1: Power on physically hosts/nodes
Option 1.1. We can power on the Nutanix AHV nodes / host physically press the power button in front of the node
Option 1.2. We can power on the Nutanix AHV nodes to browser IPMI interface to power it on
Before the start the nutanix AHV cluster need to wait 10-15 minute to boot up the all servers ( included AHV host and CVM) and ping all AHV host and CVMs.
Note : CVM boots automatically after AHV hosts booted successfully.
Step 2: SSH in to any CVM start cluster by issuing following command
cvm$ cluster start
To check cluster services are running issue following command
cvm$ cluster status
Step 3: Now do power on all guest VMs from Prism or command line
Option 3.1 Open Prism > Home > VM > select VM > click on “Power on”
Option 3.2 From CLI
SSH in to any CVM to power on all guest VMs by issuing following single command:
cvm$ acli vm.on *
Step 4: Log in to Prism > VM menu > VM table
Check your all VMs should be power on.
- Top 10 Nutanix Acropolis AHV Commands – Part 1
- Nutanix Acropolis AHV Core Architecture Explained
- Nutanix Cluster size Limitation, Scabalibity or Maximums
- How To Change Nutanix CVM, AHV and IPMI Passwords
To shutdown or power off the Nutanix Acropolis AHV cluster has best practice to shutdown the Nutanix cluster.
#Enjoy Nutanix with HyperHCI 🙂