You shut down the components of a VI workload domain that runs containerized workloads in VMware Cloud Foundation in a specific order to keep components operational by maintaining the necessary infrastructure, networking, and management services as long as possible before shutdown.

You shut down the management components for the VI workload domains that run vSphere with Tanzu and containers or that run virtualized workloads before you shut down the components for the management domain.

If the NSX Manager cluster and NSX Edge cluster are shared with other VI workload domains, follow this general order:

  1. Shut down the customer workloads in all VI workload domains that share the NSX-T Data Center instance. Otherwise, all NSX networking services in the customer workloads will be interrupted when you shut down NSX-T Data Center.

  2. Shut down the VI workload domain that runs the shared NSX Edge nodes.

  3. Shut down the other VI workload domains.

Shutdown Order for a VI Workload Domain with vSphere with Tanzu

Table 1. Shutdown Order for a VI Workload Domain with vSphere with Tanzu

Shutdown Order

SDDC Component

1

Containerized customer workloads

2

Find out the location of the vSphere with Tanzu virtual machines *

3

vSphere Cluster Services virtual machines in the VI workload domain *

4

vCenter Server for the VI workload domain *

5

Supervisor Cluster Control Plane virtual machines

6

Tanzu Kubernetes cluster control plane virtual machines

7

Tanzu Kubernetes cluster worker virtual machines

8

Harbor virtual machines

9

NSX Edge nodes in the VI workload domain *

10

NSX Manager nodes for the VI workload domain *

11

vSAN and ESXi hosts in the VI workload domain *

* For information on the shutdown steps, see this article.

Find Out the Location of the vSphere with Tanzu Virtual Machines on the ESXi Hosts

Before you begin shutting down a VI workload domain with vSphere with Tanzu, you get a mapping between virtual machines in the workload domain and the ESXi hosts on which they are deployed. You later use this mapping to log in to specific ESXi hosts and shut down specific management virtual machines.

Procedure

  1. Start Windows PowerShell.
  2. Connect to the VI workload domain vCenter Server by running the command. 
    Connect-VIServer -Server <workload_domain_vCenter_server_fqdn> -User [email protected] -Password vsphere_admin_password
  3. Generate the virtual machine to host mapping in a C:\VMToHostMapping.csv file on the Windows machine by running the command. 
    Get-VM | Select Name,VMHost | Export-Csv -Path C:\VMToHostMapping.csv -NoTypeInformation

Shut Down the vSphere Cluster Services Virtual Machines

To shut down the vSphere Cluster Services (vCLS) virtual machines in a cluster in the management domain or in a VI workload domain in VMware Cloud Foundation, you put the cluster in retreat mode. The retreat mode triggers clean-up of the vCLS virtual machines.

Procedure

  1. Log in to vCenter Server for the management or VI workload domain at https://<vcenter_server_fqdn>/ui as [email protected].
  2. In the Hosts and clusters inventory, expand the tree of the VI workload domain vCenter Server and expand the data center for the VI workload domain.
  3. Select the cluster on which vCLS must be shut down.
  4. Copy the cluster domain ID domain-c(cluster_domain_id) from the URL of the browser.

    When you navigate to a cluster in the vSphere client, the URL is similar to this one:

    https://<fqdn-of-vCenter-server>/ui/app/cluster;nav=h/urn:vmomi:ClusterComputeResource:domain-c8:eef257af-fa50-455a-af7a-6899324fabe6/summary

    You copy only domain-c8.

  5. In the Host and Clusters inventory, select the vCenter Server instance for the management domain or the VI workload domain and click the Configure tab.
  6. Under Advanced Settings, click the Edit Settings button.
  7. Locate the config.vcls.clusters.domain-c(number).enabled property for the domain cluster ID from 4 and set it to false.

    If the property is not present, add it. The entry for the cluster cannot be deleted from the vSphere Client then. However, keeping this entry is not an issue.

  8. Click Save.

Results

The vCLS monitoring service initiates the clean-up of vCLS VMs. If vSphere DRS is enabled for the cluster, it stops working and you see an additional warning in the cluster summary. vSphere DRS remains disabled until vCLS is re-enabled on this cluster.

Shut Down vCenter Server for a Virtual Infrastructure Workload Domain with vSphere With Tanzu

To shut down the vCenter Server instance for a VI workload domain with vSphere with Tanzu in VMware Cloud Foundation, you use the vSphere Client. You stop the Kubernetes services and check the vSAN health status.

Procedure

  1. Shut down the Kubertenes services on vCenter Server.
    1. Log in to vCenter Server as root by using a Secure Shell (SSH) client.
    2. To switch to the Bash shell, run the shell command.
    3. Stop the Kubernetes services by running the command. 
      vmon-cli -k wcp
    4. Verify the Kubernetes services status by running the command. 
      vmon-cli -s wcp

      The output must contain RunState: STOPPED.

  2. Log in to the management domain vCenter Server at https://<vcenter_server_fqdn>/ui as [email protected].
  3. Verify the vSAN health and resynchronization status.
    1. Select the vSAN cluster and click the Monitor tab.
    2. In the left pane, navigate to vSAN > Skyline health and verify the status of each vSAN health check category.
    3. In the left pane, under vSAN > Resyncing objects, verify that all synchronization tasks are complete.
  4. If a vSAN cluster in the workload domain has vSphere HA turned on, stop vSphere HA to avoid vSphere HA initiated migrations of virtual machines after vSAN is partitioned during the shutdown process.
    1. Select the vSAN cluster and click the Configure tab.
    2. In the left pane, select Services > vSphere Availability and click the Edit button.
    3. In the Edit Cluster Settings dialog box, turn off vSphere HA and click OK.

    This operation takes several minutes to complete.

  5. In the VMs and templates inventory, expand the management domain vCenter Server tree and expand the management domain data center.
  6. Shut down vCenter Server.
    1. Locate the vCenter Server virtual machine for the VI workload domain.
    2. Right-click the virtual machine and select Power > Shut down Guest OS.
    3. In the confirmation dialog box, click Yes.

    This operation takes several minutes to complete.

Shut Down the NSX Edge Nodes for vSphere with Tanzu

You begin shutting down the NSX-T Data Center infrastructure in a VI workload domain with vSphere with Tanzu by shutting down the NSX Edge nodes that provide north-south traffic connectivity between the physical data center networks and the NSX SDN networks.

Because the vCenter Server instance for the domain is already down, you shut down the NSX Edge nodes from the ESXi hosts where they are running.

Procedure

  1. Log in to the ESXi host that runs the first NSX Edge node as root by using the VMware Host Client.
  2. In the navigation pane, click Virtual machines.
  3. Right-click an NSX Edge virtual machine, and select Guest OS > Shut down
  4. In the confirmation dialog box, click Yes.
  5. Repeat these steps to shut down the remaining NSX Edge nodes for the VI workload domain with vSphere with Tanzu.

Shut Down the NSX Manager Nodes

You continue shutting down the NSX-T Data Center infrastructure in the management domain and a VI workload domain by shutting down the three-node NSX Manager cluster by using the vSphere Client.

Procedure

  1. Log in to the management domain vCenter Server at https://<vcenter_server_fqdn>/ui as [email protected].
  2. In the VMs and templates inventory, expand the management domain vCenter Server tree and expand the management domain data center.
  3. Right-click the primary NSX manager virtual machine and select Power > Shut down Guest OS.
  4. In the confirmation dialog box, click Yes.

    This operation takes several minutes to complete.

  5. Repeat the steps for the remaining NSX Manager virtual machines.

Shut Down vSAN and the ESXi Hosts in the Management Domain or for vSphere with Tanzu

You shut down vSAN and the ESXi hosts in the management domain or in a VI workload domain with vSphere with Tanzu by preparing the vSAN cluster for shutdown, placing each ESXi host in maintenance mode to prevent any virtual machines being deployed to or starting up on the host, and shutting down the host.

In a VI workload domain with vSphere with Tanzu, the vCenter Server instance for the domain is already down. Hence, you perform the shutdown operation on the ESXi hosts by using the VMware Host Client.

Procedure

  1. Log in to the first ESXi host for the management domain or the VI workload domain cluster by using a Secure Shell (SSH) client as root.
  2. For a vSAN cluster, disable vSAN cluster member updates by running the command. 
    esxcfg-advcfg -s 1 /VSAN/IgnoreClusterMemberListUpdates

    The command returns Value of IgnoreClusterMemberListUpdates is 1

  3. Repeat Step 1 and Step 2 on the remaining hosts in the management domain or the VI workload domain cluster.
  4. On the first ESXi host per vSAN cluster, prepare the vSAN cluster for shutdown by running the command. 
    python /usr/lib/vmware/vsan/bin/reboot_helper.py prepare

    The command returns Cluster preparation is done!

  5. Place the ESXi host in maintenance mode by running the command. 
    esxcli system maintenanceMode set -e true -m noAction

    Ensure the prompt comes back after the command is complete.

  6. Verify that the host is in maintenance mode. 
    esxcli system maintenanceMode get
  7. Repeat Step 5 and Step 6 on the remaining hosts in the management domain or VI workload domain cluster, proceeding to the next host after the operation on the current is complete.
  8. Shut down the ESXi host.
    1. Log in to the first ESXi host for the workload domain at https://<esxi_host_fqdn>/ui as root.
    2. In the navigation pane, right-click Host and, from the drop-down menu, select Shut down.
    3. In the confirmation dialog box, click Shut down.
    4. Repeat the steps on the remaining ESXi hosts in the management domain or in the VI workload domain cluster.