Deploy Management Clusters from a Configuration File

You can use the Tanzu CLI to deploy a management cluster to vSphere with a configuration that you specify in a YAML configuration file. To create a cluster configuration file, you can copy an existing configuration file for a previous deployment to vSphere and update it. Alternatively, you can create a file from scratch by using an empty template.

About Creating a Management Cluster Configuration File

When you deploy a management cluster from the CLI, you specify your cluster configuration file by using the --file option of the tanzu mc create command.

If you have previously deployed a management cluster by running tanzu mc create --ui, the ~/.config/tanzu/tkg/clusterconfigs directory contains management cluster configuration files with settings saved from each invocation of the installer interface. You can use these files as templates for cluster configuration files for new deployments. Alternatively, you can create management cluster configuration files from the template that is provided in this topic.

VMware recommends using a dedicated configuration file for each management cluster.

Important

  • As described in Configuring the Management Cluster, environment variables override values from a cluster configuration file. To use all settings from a cluster configuration file, unset any conflicting environment variables before you deploy the management cluster from the CLI.
  • Support for IPv6 addresses in Tanzu Kubernetes Grid is limited; see Deploy Clusters on IPv6. If you are not deploying to an IPv6-only networking environment, all IP address settings in your configuration files must be IPv4.
  • Some parameters configure identical properties. For example, the SIZE property configures the same infrastructure settings as all of the control plane and worker node size and type properties for the different target platforms, but at a more general level. In such cases, avoid setting conflicting or redundant properties.

Prerequisites

Before you can deploy a management cluster, you must make sure that your environment meets the requirements for the target platform.

General Prerequisites

  • Make sure that you have met the all of the requirements and followed all of the procedures in Install the Tanzu CLI and Kubernetes CLI for Use with Standalone Management Clusters.
  • For production deployments, it is strongly recommended to enable identity management for your clusters. For information about the preparatory steps to perform before you deploy a management cluster, see Obtain Your Identity Provider Details in Configure Identity Management. For conceptual information about identity management and access control in Tanzu Kubernetes Grid, see About Identity and Access Management.
  • If you are deploying clusters in an internet-restricted environment, you must also perform the steps in Prepare an Internet-Restricted Environment. These steps include setting TKG_CUSTOM_IMAGE_REPOSITORY as an environment variable.
  • Important

    It is strongly recommended to use the Tanzu Kubernetes Grid installer interface rather than the CLI to deploy your first management cluster. When you deploy a management cluster by using the installer interface, it populates a cluster configuration file for the management cluster with the required parameters. You can use the created configuration file as a model for future deployments from the CLI to your target platform.

  • If you plan on registering the management cluster with Tanzu Mission Control, ensure that your workload clusters meet the requirements listed in Requirements for Registering a Tanzu Kubernetes Cluster with Tanzu Mission Control in the Tanzu Mission Control documentation.

Infrastructure Prerequisites

Make sure that you have met the all of the requirements listed in Prepare to Deploy Management Clusters to vSphere.

Important

On vSphere with Tanzu, you may not need to deploy a management cluster. See vSphere with Tanzu Supervisor is a Management Cluster.

Create a Management Cluster Configuration File

To create a configuration file for a standalone management cluster by using a template:

  1. In a text editor, open a new file with a .yaml extension and an appropriate name, for example, vsphere-mgmt-cluster-config.yaml. This will be your configuration file.

    If you have already deployed a management cluster from the installer interface, you can create the file in the default location for cluster configurations, ~/.config/tanzu/tkg/clusterconfigs.

  2. Copy-paste the Management Cluster Configuration Template code into your configuration file.

  3. Configure settings within the file by following the instructions in this topic. For information about all of the variables that the configuration file can include, see the Configuration File Variable Reference.

  4. Save the file.

To use the configuration file from a previous deployment that you performed by using the installer interface, make a copy of the configuration file with a new name, open it in a text editor, and update the configuration. For information about how to update all of the settings, see the Configuration File Variable Reference.

Management Cluster Configuration Template

The template below includes all of the options that are relevant to deploying management clusters on vSphere. You can copy this template and use it to deploy management clusters to vSphere.

  • For information about all configuration file variables, see the Configuration File Variable Reference.
  • For examples of how to configure these settings, see the sections below the template.

Mandatory options are uncommented. Optional settings are commented out. Default values are included where applicable.

#! ---------------------------------------------------------------------
#! Basic cluster creation configuration
#! ---------------------------------------------------------------------

CLUSTER_NAME:
CLUSTER_PLAN: dev
INFRASTRUCTURE_PROVIDER: vsphere
# CLUSTER_API_SERVER_PORT: # For deployments without NSX Advanced Load Balancer
ENABLE_CEIP_PARTICIPATION: true
ENABLE_AUDIT_LOGGING: true
CLUSTER_CIDR: 100.96.0.0/11
SERVICE_CIDR: 100.64.0.0/13
# CAPBK_BOOTSTRAP_TOKEN_TTL: 30m

#! ---------------------------------------------------------------------
#! vSphere configuration
#! ---------------------------------------------------------------------

VSPHERE_SERVER:
VSPHERE_USERNAME:
VSPHERE_PASSWORD:
VSPHERE_DATACENTER:
VSPHERE_RESOURCE_POOL:
VSPHERE_DATASTORE:
VSPHERE_FOLDER:
VSPHERE_NETWORK: VM Network
# VSPHERE_CONTROL_PLANE_ENDPOINT: # Required for Kube-Vip
# VSPHERE_CONTROL_PLANE_ENDPOINT_PORT: 6443
VIP_NETWORK_INTERFACE: "eth0"
# VSPHERE_TEMPLATE:
VSPHERE_SSH_AUTHORIZED_KEY:
# VSPHERE_STORAGE_POLICY_ID: ""
VSPHERE_TLS_THUMBPRINT:
VSPHERE_INSECURE: false
DEPLOY_TKG_ON_VSPHERE7: false
ENABLE_TKGS_ON_VSPHERE7: false

#! ---------------------------------------------------------------------
#! Node configuration
#! ---------------------------------------------------------------------

# SIZE:
# CONTROLPLANE_SIZE:
# WORKER_SIZE:
# OS_NAME: ""
# OS_VERSION: ""
# OS_ARCH: ""
# VSPHERE_NUM_CPUS: 2
# VSPHERE_DISK_GIB: 40
# VSPHERE_MEM_MIB: 4096
# VSPHERE_MTU:
# VSPHERE_CONTROL_PLANE_NUM_CPUS: 2
# VSPHERE_CONTROL_PLANE_DISK_GIB: 40
# VSPHERE_CONTROL_PLANE_MEM_MIB: 8192
# VSPHERE_WORKER_NUM_CPUS: 2
# VSPHERE_WORKER_DISK_GIB: 40
# VSPHERE_WORKER_MEM_MIB: 4096

#! ---------------------------------------------------------------------
#! VMware NSX specific configuration for enabling NSX routable pods
#! ---------------------------------------------------------------------

# NSXT_POD_ROUTING_ENABLED: false
# NSXT_ROUTER_PATH: ""
# NSXT_USERNAME: ""
# NSXT_PASSWORD: ""
# NSXT_MANAGER_HOST: ""
# NSXT_ALLOW_UNVERIFIED_SSL: false
# NSXT_REMOTE_AUTH: false
# NSXT_VMC_ACCESS_TOKEN: ""
# NSXT_VMC_AUTH_HOST: ""
# NSXT_CLIENT_CERT_KEY_DATA: ""
# NSXT_CLIENT_CERT_DATA: ""
# NSXT_ROOT_CA_DATA: ""
# NSXT_SECRET_NAME: "cloud-provider-vsphere-nsxt-credentials"
# NSXT_SECRET_NAMESPACE: "kube-system"

#! ---------------------------------------------------------------------
#! NSX Advanced Load Balancer configuration
#! ---------------------------------------------------------------------

AVI_ENABLE: false
AVI_CONTROL_PLANE_HA_PROVIDER: false
# AVI_NAMESPACE: "tkg-system-networking"
# AVI_DISABLE_INGRESS_CLASS: true
# AVI_AKO_IMAGE_PULL_POLICY: IfNotPresent
# AVI_ADMIN_CREDENTIAL_NAME: avi-controller-credentials
# AVI_CA_NAME: avi-controller-ca
# AVI_CONTROLLER:
# AVI_USERNAME: ""
# AVI_PASSWORD: ""
# AVI_CLOUD_NAME:
# AVI_SERVICE_ENGINE_GROUP:
# AVI_NSXT_T1LR: # Required for NSX ALB deployments on NSX Cloud.
# AVI_MANAGEMENT_CLUSTER_SERVICE_ENGINE_GROUP:
# AVI_DATA_NETWORK:
# AVI_DATA_NETWORK_CIDR:
# AVI_MANAGEMENT_CLUSTER_VIP_NETWORK_NAME:
# AVI_MANAGEMENT_CLUSTER_VIP_NETWORK_CIDR:
# AVI_CA_DATA_B64: ""
# AVI_LABELS: ""
# AVI_DISABLE_STATIC_ROUTE_SYNC: true
# AVI_INGRESS_DEFAULT_INGRESS_CONTROLLER: false
# AVI_INGRESS_SHARD_VS_SIZE: ""
# AVI_INGRESS_SERVICE_TYPE: ""
# AVI_INGRESS_NODE_NETWORK_LIST: ""

#! ---------------------------------------------------------------------
#! Image repository configuration
#! ---------------------------------------------------------------------

# TKG_CUSTOM_IMAGE_REPOSITORY: ""
# TKG_CUSTOM_IMAGE_REPOSITORY_CA_CERTIFICATE: ""

#! ---------------------------------------------------------------------
#! Proxy configuration
#! ---------------------------------------------------------------------

# TKG_HTTP_PROXY: ""
# TKG_HTTPS_PROXY: ""
# TKG_NO_PROXY: ""

#! ---------------------------------------------------------------------
#! Machine Health Check configuration
#! ---------------------------------------------------------------------

ENABLE_MHC:
ENABLE_MHC_CONTROL_PLANE: true
ENABLE_MHC_WORKER_NODE: true
MHC_UNKNOWN_STATUS_TIMEOUT: 5m
MHC_FALSE_STATUS_TIMEOUT: 12m

#! ---------------------------------------------------------------------
#! Identity management configuration
#! ---------------------------------------------------------------------

IDENTITY_MANAGEMENT_TYPE: "none"

#! Settings for IDENTITY_MANAGEMENT_TYPE: "oidc"
# CERT_DURATION: 2160h
# CERT_RENEW_BEFORE: 360h
# OIDC_IDENTITY_PROVIDER_CLIENT_ID:
# OIDC_IDENTITY_PROVIDER_CLIENT_SECRET:
# OIDC_IDENTITY_PROVIDER_GROUPS_CLAIM: groups
# OIDC_IDENTITY_PROVIDER_ISSUER_URL:
# OIDC_IDENTITY_PROVIDER_SCOPES: "email,profile,groups,offline_access"
# OIDC_IDENTITY_PROVIDER_USERNAME_CLAIM: email

#! The following two variables are used to configure Pinniped JWTAuthenticator for workload clusters
# SUPERVISOR_ISSUER_URL:
# SUPERVISOR_ISSUER_CA_BUNDLE_DATA:

#! Settings for IDENTITY_MANAGEMENT_TYPE: "ldap"
# LDAP_BIND_DN:
# LDAP_BIND_PASSWORD:
# LDAP_HOST:
# LDAP_USER_SEARCH_BASE_DN:
# LDAP_USER_SEARCH_FILTER:
# LDAP_USER_SEARCH_ID_ATTRIBUTE: dn
# LDAP_USER_SEARCH_NAME_ATTRIBUTE:
# LDAP_GROUP_SEARCH_BASE_DN:
# LDAP_GROUP_SEARCH_FILTER:
# LDAP_GROUP_SEARCH_NAME_ATTRIBUTE: dn
# LDAP_GROUP_SEARCH_USER_ATTRIBUTE: dn
# LDAP_ROOT_CA_DATA_B64:

#! ---------------------------------------------------------------------
#! Antrea CNI configuration
#! ---------------------------------------------------------------------

# ANTREA_NO_SNAT: true
# ANTREA_NODEPORTLOCAL: true
# ANTREA_NODEPORTLOCAL_ENABLED: true
# ANTREA_NODEPORTLOCAL_PORTRANGE: 61000-62000
# ANTREA_TRAFFIC_ENCAP_MODE: "encap"
# ANTREA_PROXY: true
# ANTREA_PROXY_ALL: false
# ANTREA_PROXY_LOAD_BALANCER_IPS: false
# ANTREA_PROXY_NODEPORT_ADDRS:
# ANTREA_PROXY_SKIP_SERVICES: ""
# ANTREA_POLICY: true
# ANTREA_TRACEFLOW: true
# ANTREA_DISABLE_UDP_TUNNEL_OFFLOAD: false
# ANTREA_ENABLE_USAGE_REPORTING: false
# ANTREA_EGRESS: true
# ANTREA_EGRESS_EXCEPT_CIDRS: ""
# ANTREA_FLOWEXPORTER: false
# ANTREA_FLOWEXPORTER_COLLECTOR_ADDRESS: "flow-aggregator.flow-aggregator.svc:4739:tls"
# ANTREA_FLOWEXPORTER_POLL_INTERVAL: "5s"
# ANTREA_FLOWEXPORTER_ACTIVE_TIMEOUT: "5s"
# ANTREA_FLOWEXPORTER_IDLE_TIMEOUT: "15s"
# ANTREA_IPAM: false
# ANTREA_KUBE_APISERVER_OVERRIDE: ""
# ANTREA_MULTICAST: false
# ANTREA_MULTICAST_INTERFACES: ""
# ANTREA_NETWORKPOLICY_STATS: true
# ANTREA_SERVICE_EXTERNALIP: true
# ANTREA_TRANSPORT_INTERFACE: ""
# ANTREA_TRANSPORT_INTERFACE_CIDRS: ""

Configure Settings in the Management Cluster Configuration File

Follow the instructions below to configure your management cluster deployment.

Configure Basic Management Cluster Creation Information

The basic management cluster creation settings define the infrastructure on which to deploy the management cluster and other basic settings.

  • For CLUSTER_PLAN specify whether you want to deploy a development cluster, which provides a single control plane node, or a production cluster, which provides a highly available management cluster with three control plane nodes. Specify dev or prod.
  • For INFRASTRUCTURE_PROVIDER, specify vsphere.
  • Optionally, deactivate participation in the VMware Customer Experience Improvement Program (CEIP) by setting ENABLE_CEIP_PARTICIPATION to false. For information about the CEIP, see Manage Participation in CEIP and https://www.vmware.com/solutions/trustvmware/ceip.html.
  • Optionally, deactivate audit logging by setting ENABLE_AUDIT_LOGGING to false. For information about audit logging, see Audit Logging.
  • If the recommended CIDR ranges of 100.64.0.0/13 and 100.96.0.0/11 are unavailable, update CLUSTER_CIDR for the cluster pod network and SERVICE_CIDR for the cluster service network.

For example:

#! ---------------------------------------------------------------------
#! Basic cluster creation configuration
#! ---------------------------------------------------------------------

CLUSTER_NAME: vsphere-mgmt-cluster
CLUSTER_PLAN: dev
INFRASTRUCTURE_PROVIDER: vsphere
ENABLE_CEIP_PARTICIPATION: true
ENABLE_AUDIT_LOGGING: true
CLUSTER_CIDR: 100.96.0.0/11
SERVICE_CIDR: 100.64.0.0/13

Configure General vSphere Settings

Provide information to allow Tanzu Kubernetes Grid to log in to vSphere and to designate the resources that Tanzu Kubernetes Grid can use.

  • Update the VSPHERE_SERVER, VSPHERE_USERNAME, and VSPHERE_PASSWORD settings with the IP address or FQDN of the vCenter Server instance and the credentials to use to log in.
  • Provide the full paths to the vSphere datacenter, resource pool, datastores, and folder in which to deploy the management cluster:

    • VSPHERE_DATACENTER: /<MY-DATACENTER>
    • VSPHERE_RESOURCE_POOL: /<MY-DATACENTER>/host/<CLUSTER>/Resources
    • VSPHERE_DATASTORE: /<MY-DATACENTER>/datastore/<MY-DATASTORE>
    • VSPHERE_FOLDER: /<MY-DATACENTER>/vm/<FOLDER>.
  • Depending on the HA provider for the cluster’s control plane API, set VSPHERE_CONTROL_PLANE_ENDPOINT or leave it blank:
    • Kube-VIP: Set to a static virtual IP address, or to a fully qualified domain name (FQDN) mapped to the VIP address.
    • NSX Advanced Load Balancer: Leave blank, unless you need to specify an endpoint. If so, use a static address within the IPAM Profile’s VIP Network range that you have manually added to the Static IP pool.
  • Specify a network and a network interface in VSPHERE_NETWORK and VIP_NETWORK_INTERFACE.
  • Optionally, uncomment and update VSPHERE_TEMPLATE to specify the path to an OVA file if you are using multiple custom OVA images for the same Kubernetes version. Use the format /MY-DC/vm/MY-FOLDER-PATH/MY-IMAGE. For more information, see Deploy a Cluster with a Custom OVA Image in Creating and Managing TKG 2.5 Workload Clusters on vSphere with the Tanzu CLI.
  • Provide your SSH key in the VSPHERE_SSH_AUTHORIZED_KEY option. For information about how to obtain an SSH key, see Prepare to Deploy Management Clusters to vSphere.
  • Provide the TLS thumbprint in the VSPHERE_TLS_THUMBPRINT variable, or set VSPHERE_INSECURE: true to skip thumbprint verification.
  • Optionally, uncomment VSPHERE_STORAGE_POLICY_ID and specify the name of a storage policy for the VMs, which you have configured on vCenter Server, for the management cluster to use.

For example:

#! ---------------------------------------------------------------------
#! vSphere configuration
#! ---------------------------------------------------------------------

VSPHERE_SERVER: 10.185.12.154
VSPHERE_USERNAME: [email protected]
VSPHERE_PASSWORD: <encoded:QWRtaW4hMjM=>
VSPHERE_DATACENTER: /dc0
VSPHERE_RESOURCE_POOL: /dc0/host/cluster0/Resources/tanzu
VSPHERE_DATASTORE: /dc0/datastore/sharedVmfs-1
VSPHERE_FOLDER: /dc0/vm/tanzu
VSPHERE_NETWORK: "VM Network"
VSPHERE_CONTROL_PLANE_ENDPOINT: 10.185.11.134
VIP_NETWORK_INTERFACE: "eth0"
VSPHERE_TEMPLATE: /dc0/vm/tanzu/my-image.ova
VSPHERE_SSH_AUTHORIZED_KEY: ssh-rsa AAAAB3[...]tyaw== [email protected]
VSPHERE_TLS_THUMBPRINT: 47:F5:83:8E:5D:36:[...]:72:5A:89:7D:29:E5:DA
VSPHERE_INSECURE: false
VSPHERE_STORAGE_POLICY_ID: "My storage policy"

Configure Nodes

The Tanzu CLI creates the individual nodes of management clusters and workload clusters according to settings that you provide in the configuration file. You can configure all node VMs to have the same predefined configurations, set different predefined configurations for control plane and worker nodes, or customize the configurations of the nodes. By using these settings, you can create clusters that have nodes with different configurations to the management cluster nodes. You can also create clusters in which the control plane nodes and worker nodes have different configurations.

Predefined Node Sizes

The Tanzu CLI provides the following predefined configurations for cluster nodes:

  • small: 2 CPUs, 4 GB memory, 20 GB disk
  • medium: 2 CPUs, 8 GB memory, 40 GB disk
  • large: 4 CPUs, 16 GB memory, 40 GB disk
  • extra-large: 8 CPUs, 32 GB memory, 80 GB disk

To create a cluster in which all of the control plane and worker node VMs are the same size, specify the SIZE variable. If you set the SIZE variable, all nodes will be created with the configuration that you set.

SIZE: "large"

To create a in which the control plane and worker node VMs are different sizes, specify the CONTROLPLANE_SIZE and WORKER_SIZE options.

CONTROLPLANE_SIZE: "medium"
WORKER_SIZE: "extra-large"

You can combine the CONTROLPLANE_SIZE and WORKER_SIZE options with the SIZE option. For example, if you specify SIZE: "large" with WORKER_SIZE: "extra-large", the control plane nodes will be set to large and worker nodes will be set to extra-large.

SIZE: "large"
WORKER_SIZE: "extra-large"

Custom Node Sizes

You can customize the configuration of the nodes rather than using the predefined configurations.

To use the same custom configuration for all nodes, specify the VSPHERE_NUM_CPUS, VSPHERE_DISK_GIB, and VSPHERE_MEM_MIB options.

VSPHERE_NUM_CPUS: 2
VSPHERE_DISK_GIB: 40
VSPHERE_MEM_MIB: 4096

To define different custom configurations for control plane nodes and worker nodes, specify the VSPHERE_CONTROL_PLANE_* and VSPHERE_WORKER_* options.

VSPHERE_CONTROL_PLANE_NUM_CPUS: 2
VSPHERE_CONTROL_PLANE_DISK_GIB: 20
VSPHERE_CONTROL_PLANE_MEM_MIB: 8192
VSPHERE_WORKER_NUM_CPUS: 4
VSPHERE_WORKER_DISK_GIB: 40
VSPHERE_WORKER_MEM_MIB: 4096

You can override these settings by using the SIZE, CONTROLPLANE_SIZE, and WORKER_SIZE options.

OS Settings

By default, all cluster nodes run Ubuntu v22.04. You can optionally deploy clusters that run Photon OS on their nodes. For the architecture, the default and only current choice is amd64. For the OS and version settings, see see Node Configuration in the Configuration File Variable Reference.

For example:

#! ---------------------------------------------------------------------
#! Node configuration
#! ---------------------------------------------------------------------

OS_NAME: "photon"
OS_VERSION: "5"
OS_ARCH: "amd64"

Configure NSX Routable Pods

If your vSphere environment uses NSX, you can configure it to implement routable, or NO_NAT, pods.

Note

NSX Routable Pods is an experimental feature in this release. Information about how to implement NSX Routable Pods will be added to this documentation soon.

#! ---------------------------------------------------------------------
#! NSX specific configuration for enabling NSX routable pods
#! ---------------------------------------------------------------------

# NSXT_POD_ROUTING_ENABLED: false
# NSXT_ROUTER_PATH: ""
# NSXT_USERNAME: ""
# NSXT_PASSWORD: ""
# NSXT_MANAGER_HOST: ""
# NSXT_ALLOW_UNVERIFIED_SSL: false
# NSXT_REMOTE_AUTH: false
# NSXT_VMC_ACCESS_TOKEN: ""
# NSXT_VMC_AUTH_HOST: ""
# NSXT_CLIENT_CERT_KEY_DATA: ""
# NSXT_CLIENT_CERT_DATA: ""
# NSXT_ROOT_CA_DATA: ""
# NSXT_SECRET_NAME: "cloud-provider-vsphere-nsxt-credentials"
# NSXT_SECRET_NAMESPACE: "kube-system"

Configure NSX Advanced Load Balancer

To use NSX Advanced Load Balancer, you must first deploy it in your vSphere environment. See Install NSX Advanced Load Balancer. After deploying NSX Advanced Load Balancer, configure a vSphere management cluster to use the load balancer.

For example:

AVI_ENABLE: true
AVI_CONTROL_PLANE_HA_PROVIDER: true
AVI_NAMESPACE: "tkg-system-networking"
AVI_DISABLE_INGRESS_CLASS: true
AVI_AKO_IMAGE_PULL_POLICY: IfNotPresent
AVI_ADMIN_CREDENTIAL_NAME: avi-controller-credentials
AVI_CA_NAME: avi-controller-ca
AVI_CONTROLLER: 10.185.10.217
AVI_USERNAME: "admin"
AVI_PASSWORD: "<password>"
AVI_CLOUD_NAME: "Default-Cloud"
AVI_SERVICE_ENGINE_GROUP: "Default-Group"
AVI_NSXT_T1LR:""
AVI_DATA_NETWORK: nsx-alb-dvswitch
AVI_DATA_NETWORK_CIDR: 10.185.0.0/20
AVI_MANAGEMENT_CLUSTER_VIP_NETWORK_NAME: ""
AVI_MANAGEMENT_CLUSTER_VIP_NETWORK_CIDR: ""
AVI_CA_DATA_B64: LS0tLS1CRU[...]UtLS0tLQo=
AVI_LABELS: ""
AVI_DISABLE_STATIC_ROUTE_SYNC: true
AVI_INGRESS_DEFAULT_INGRESS_CONTROLLER: false
AVI_INGRESS_SHARD_VS_SIZE: ""
AVI_INGRESS_SERVICE_TYPE: ""
AVI_INGRESS_NODE_NETWORK_LIST: ""

By default, the management cluster and all workload clusters that it manages will use the load balancer. For information about how to configure the NSX Advanced Load Balancer variables, see NSX Advanced Load Balancer in the Configuration File Variable Reference.

NSX Advanced Load Balancer as a Control Plane Endpoint Provider

You can use NSX ALB as the control plane endpoint provider in Tanzu Kubernetes Grid. The following table describes the differences between NSX ALB and Kube-Vip, which is the default control plane endpoint provider in Tanzu Kubernetes Grid.

Kube-Vip NSX ALB
Sends Traffic to Single control plane node
Multiple control plane nodes
Requires configuring endpoint VIP Yes
No
Assigns VIP from the NSX ALB static IP pool

Configure a Private Image Registry

If you are deploying the management cluster in an Internet-restricted environment, uncomment and update the TKG_CUSTOM_IMAGE_REPOSITORY_* settings. You do not need to configure the private image registry settings if:

  • You are deploying the management cluster in an internet-restricted environment and you have set the TKG_CUSTOM_IMAGE_REPOSITORY_* variables by running the tanzu config set command, as described in Prepare an Internet-Restricted Environment. Environment variables set by running tanzu config set override values from a cluster configuration file.
  • If you are deploying the management cluster in an environment that has access to the external internet.

For example:

#! ---------------------------------------------------------------------
#! Image repository configuration
#! ---------------------------------------------------------------------

TKG_CUSTOM_IMAGE_REPOSITORY: "custom-image-repository.io/yourproject"
TKG_CUSTOM_IMAGE_REPOSITORY_CA_CERTIFICATE: "LS0t[...]tLS0tLQ=="

Configure Proxies

To optionally send outgoing HTTP(S) traffic from the management cluster to a proxy, for example, in an internet-restricted environment, uncomment and set the *_PROXY settings. You can choose to use one proxy for HTTP requests and another proxy for HTTPS requests or to use the same proxy for both HTTP and HTTPS requests. You cannot change the proxy after you deploy the cluster.

Note

Traffic from cluster VMs to vCenter cannot be proxied. In a proxied vSphere environment, you need to either set VSPHERE_INSECURE to true, or else add the vCenter IP address or hostname to the TKG_NO_PROXY list.

  • TKG_HTTP_PROXY_ENABLED: Set this to true to configure a proxy.

  • TKG_PROXY_CA_CERT: Set this to the proxy server’s CA if its certificate is self-signed.

  • TKG_HTTP_PROXY: This is the URL of the proxy that handles HTTP requests. To set the URL, use the format below:

    PROTOCOL://USERNAME:PASSWORD@FQDN-OR-IP:PORT
    

    Where:

    • (Required) PROTOCOL: This must be http.
    • (Optional) USERNAME and PASSWORD: This is your HTTP proxy username and password. You must set USERNAME and PASSWORD if the proxy requires authentication.

    Note: When deploying management clusters with CLI, the following non-alphanumeric characters cannot be used in passwords: # ` ^ | / \ ? % ^ { [ ] }" < > .

    • (Required) FQDN-OR-IP: This is the FQDN or IP address of your HTTP proxy.
    • (Required) PORT: This is the port number that your HTTP proxy uses.

    For example, http://user:[email protected]:1234.

  • TKG_HTTPS_PROXY: This is the URL of the proxy that handles HTTPS requests. You can set TKG_HTTPS_PROXY to the same value as TKG_HTTP_PROXY or provide a different value. To set the value, use the URL format from the previous step, where:

    • (Required) PROTOCOL: This must be http.
    • (Optional) USERNAME and PASSWORD: This is your HTTPS proxy username and password. You must set USERNAME and PASSWORD if the proxy requires authentication.

    Note: When deploying management clusters with CLI, the following non-alphanumeric characters cannot be used in passwords: # ` ^ | / \ ? % ^ { [ ] }" < > .

    • (Required) FQDN-OR-IP: This is the FQDN or IP address of your HTTPS proxy.
    • (Required) PORT: This is the port number that your HTTPS proxy uses.

    For example, http://user:[email protected]:1234.

  • TKG_NO_PROXY: This sets one or more comma-separated network CIDRs or hostnames that must bypass the HTTP(S) proxy, for example to enable the management cluster to communicate directly with infrastructure that runs on the same network, behind the same proxy. Do not use spaces in the comma-separated list setting. For example, noproxy.yourdomain.com,192.168.0.0/24.

    This list must include:

    • Your vCenter IP address or hostname.
    • The CIDR of VSPHERE_NETWORK, which includes the IP address of your control plane endpoint. If you set VSPHERE_CONTROL_PLANE_ENDPOINT to an FQDN, also add that FQDN to the TKG_NO_PROXY list.

    Internally, Tanzu Kubernetes Grid appends localhost, 127.0.0.1, the values of CLUSTER_CIDR and SERVICE_CIDR, .svc, and .svc.cluster.local to the value that you set in TKG_NO_PROXY. You must manually add the CIDR of VSPHERE_NETWORK, which includes the IP address of your control plane endpoint, to TKG_NO_PROXY. If you set VSPHERE_CONTROL_PLANE_ENDPOINT to an FQDN, add both the FQDN and VSPHERE_NETWORK to TKG_NO_PROXY.

    Important

    If the cluster VMs need to communicate with external services and infrastructure endpoints in your Tanzu Kubernetes Grid environment, ensure that those endpoints are reachable by the proxies that you set above or add them to TKG_NO_PROXY. Depending on your environment configuration, this may include, but is not limited to:

    • Your OIDC or LDAP server
    • Harbor
    • VMware NSX
    • NSX Advanced Load Balancer

For example:

#! ---------------------------------------------------------------------
#! Proxy configuration
#! ---------------------------------------------------------------------

TKG_HTTP_PROXY_ENABLED: true
TKG_PROXY_CA_CERT: "LS0t[...]tLS0tLQ==""
TKG_HTTP_PROXY: "http://myproxy.com:1234"
TKG_HTTPS_PROXY: "http://myproxy.com:1234"
TKG_NO_PROXY: "noproxy.yourdomain.com,192.168.0.0/24"

Configure Machine Health Checks

Optionally, update variables based on your deployment preferences and using the guidelines described in the Machine Health Checks section of Configuration File Variable Reference.

For example:

ENABLE_MHC:
ENABLE_MHC_CONTROL_PLANE: true
ENABLE_MHC_WORKER_NODE: true
MHC_MAX_UNHEALTHY_CONTROL_PLANE: 60%
MHC_MAX_UNHEALTHY_WORKER_NODE: 60%
MHC_UNKNOWN_STATUS_TIMEOUT: 10m
MHC_FALSE_STATUS_TIMEOUT: 20m

Configure Identity Management

Set IDENTITY_MANAGEMENT_TYPE to ldap or oidc. Set to none or omit to deactivate identity management. It is strongly recommended to enable identity management for production deployments.

IDENTITY_MANAGEMENT_TYPE: oidc
IDENTITY_MANAGEMENT_TYPE: ldap

OIDC

To configure OIDC, update the variables below. For information about how to configure the variables, see Identity Providers - OIDC in the Configuration File Variable Reference.

For example:

OIDC_IDENTITY_PROVIDER_CLIENT_ID: 0oa2i[...]NKst4x7
OIDC_IDENTITY_PROVIDER_CLIENT_SECRET: 331!b70[...]60c_a10-72b4
OIDC_IDENTITY_PROVIDER_GROUPS_CLAIM: groups
OIDC_IDENTITY_PROVIDER_ISSUER_URL: https://dev-[...].okta.com
OIDC_IDENTITY_PROVIDER_SCOPES: openid,groups,email,offline_access
OIDC_IDENTITY_PROVIDER_USERNAME_CLAIM: email

LDAP

To configure LDAP, uncomment and update the LDAP_* variables with information about your LDAPS server. For information about how to configure the variables, see Identity Providers - LDAP in the Configuration File Variable Reference.

For example:

LDAP_BIND_DN: "cn=bind-user,ou=people,dc=example,dc=com"
LDAP_BIND_PASSWORD: "example-password"
LDAP_GROUP_SEARCH_BASE_DN: dc=example,dc=com
LDAP_GROUP_SEARCH_FILTER: &(objectClass=posixGroup)(memberUid={})
LDAP_GROUP_SEARCH_NAME_ATTRIBUTE: cn
LDAP_GROUP_SEARCH_USER_ATTRIBUTE: uid
LDAP_HOST: ldaps.example.com:636
LDAP_ROOT_CA_DATA_B64: ""
LDAP_USER_SEARCH_BASE_DN: ou=people,dc=example,dc=com
LDAP_USER_SEARCH_FILTER: &(objectClass=posixAccount)(uid={})
LDAP_USER_SEARCH_NAME_ATTRIBUTE: uid

Configure Antrea CNI

By default, clusters that you deploy with the Tanzu CLI provide in-cluster container networking with the Antrea container network interface (CNI).

With ANTREA_* configuration variables, you can optionally deactivate Source Network Address Translation (SNAT) for pod traffic, implement hybrid, noEncap, NetworkPolicyOnly traffic encapsulation modes, use proxies and network policies, and implement Traceflow.

Proxy Settings: Antrea proxy and related configurations determine which TKG components handle network traffic with service types ClusterIP, NodePort, and LoadBalancer originating from internal pods, internal nodes, and external clients:

  • ANTREA_PROXY=true and ANTREA_PROXY_ALL=false (default): AntreaProxy handles ClusterIP traffic from pods, and kube-proxy handles all service traffic from nodes and external traffic, which has service type NodePort.
  • ANTREA_PROXY=false: kube-proxy handles all service traffic from all sources; overrides settings for ANTREA_PROXY_ALL.
  • ANTREA_PROXY_ALL=true: AntreaProxy handles all service traffic from all nodes and pods.
    • If present,kube-proxy also redundantly handles all service traffic from nodes and hostNetwork pods, including Antrea components, typically before AntreaProxy does.
    • If kube-proxy has been removed, as described in Removing kube-proxy, AntreaProxy alone serves all traffic types from all sources.
      • This configuration is Experimental and unsupported.
      • While AntreaProxy provides ClusterIP services for traffic to the Kubernetes API server, it also connects to the server itself. So it is safer to give AntreaProxy its own address for kube-apiserver.
        • You configure this as ANTREA_KUBE_APISERVER_OVERRIDE in format CONTROL-PLANE-VIP:PORT. The address should be either maintained by kube-vip or a static IP for a control plane node.
        • If this value is wrong, Antrea will crash and container networking will not work.
  • LoadBalancer service:
    • To use Antrea as an in-cluster LoadBalancer solution, enable ANTREA_SERVICE_EXTERNALIP and define Antrea ExternalIPPool custom resources as described in Service of type LoadBalancer in the Antrea documentation.
    • kube-proxy cannot serve as a load balancer and needs a third-party load balancer solution for allocating and advertising LoadBalancer IP addresses.

For more information about Antrea, see the following resources:

To optionally configure these features on Antrea, uncomment and update the ANTREA_* variables. For example:

#! ---------------------------------------------------------------------
#! Antrea CNI configuration
#! ---------------------------------------------------------------------

ANTREA_NO_SNAT: true
ANTREA_NODEPORTLOCAL: true
ANTREA_NODEPORTLOCAL_ENABLED: true
ANTREA_NODEPORTLOCAL_PORTRANGE: 61000-62000
ANTREA_TRAFFIC_ENCAP_MODE: "encap"
ANTREA_PROXY: true
ANTREA_PROXY_ALL: false
ANTREA_PROXY_LOAD_BALANCER_IPS: false
ANTREA_PROXY_NODEPORT_ADDRS:
ANTREA_PROXY_SKIP_SERVICES: ""
ANTREA_POLICY: true
ANTREA_TRACEFLOW: true
ANTREA_DISABLE_UDP_TUNNEL_OFFLOAD: false
ANTREA_ENABLE_USAGE_REPORTING: false
ANTREA_EGRESS: true
ANTREA_EGRESS_EXCEPT_CIDRS: ""
ANTREA_FLOWEXPORTER: false
ANTREA_FLOWEXPORTER_COLLECTOR_ADDRESS: "flow-aggregator.flow-aggregator.svc:4739:tls"
ANTREA_FLOWEXPORTER_POLL_INTERVAL: "5s"
ANTREA_FLOWEXPORTER_ACTIVE_TIMEOUT: "5s"
ANTREA_FLOWEXPORTER_IDLE_TIMEOUT: "15s"
ANTREA_IPAM: false
ANTREA_KUBE_APISERVER_OVERRIDE: ""
ANTREA_MULTICAST: false
ANTREA_MULTICAST_INTERFACES: ""
ANTREA_NETWORKPOLICY_STATS: true
ANTREA_SERVICE_EXTERNALIP: true
ANTREA_TRAFFIC_ENCRYPTION_MODE: none
ANTREA_TRANSPORT_INTERFACE: ""
ANTREA_TRANSPORT_INTERFACE_CIDRS: ""

Configure for IPv6

To configure a management cluster that supports IPv6 to an IPv6 networking environment:

  1. Prepare the environment as described in (Optional) Set Variables and Rules for IPv6.

  2. Set the following variables in the configuration file for the management cluster.

    • Set TKG_IP_FAMILY to ipv6.
    • Set VSPHERE_CONTROL_PLANE_ENDPOINT to a static IPv6 address.
    • (Optional) Set the CLUSTER_CIDR and SERVICE_CIDR. Defaults to fd00:100:64::/48 and fd00:100:96::/108 respectively.

Configure Multiple Availability Zones

You can configure a management or workload cluster that runs nodes in multiple availability zones (AZs) as described in Running Clusters Across Multiple Availability Zones

Prerequisites

To configure a cluster with nodes deployed across multiple AZs:

  • Set VSPHERE_REGION and VSPHERE_ZONE to the region and zone tag categories, k8s-region and k8s-zone.
  • Set VSPHERE_AZ_0, VSPHERE_AZ_1, VSPHERE_AZ_2 with the names of the VsphereDeploymentZone objects where the machines need to be deployed.
    • The VsphereDeploymentZone associated with VSPHERE_AZ_0 is the VSphereFailureDomain in which the machine deployment ending with md-0 gets deployed, similarly VSPHERE_AZ_1 is the VSphereFailureDomain in which the machine deployment ending with md-1 gets deployed, and VSPHERE_AZ_2 is the VSphereFailureDomain in which the machine deployment ending with md-2 gets deployed
    • If any of the AZ configs are not defined, then that machine deployment gets deployed without any VSphereFailureDomain
  • WORKER_MACHINE_COUNT sets the total number of workers for the cluster. The total number of workers are distributed in a round-robin fashion across the number of AZs specified
  • VSPHERE_AZ_CONTROL_PLANE_MATCHING_LABELS sets key/value selector labels for the AZs that cluster control plane nodes may deploy to.
    • Set this variable if VSPHERE_REGION and VSPHERE_ZONE are set.
    • The labels must exist in the VSphereDeploymentZone resources that you create.
    • These labels let you specify all AZs in a region and an environment without having to list them individually, for example: "region=us-west-1,environment=staging".

For the full list of options that you must specify when deploying clusters to vSphere, see the Configuration File Variable Reference.

Configure Node IPAM

TKG supports in-cluster Node IPAM for standalone management clusters on vSphere and the class-based workload clusters that they manage. For more information and current limitations, see Node IPAM in Creating and Managing TKG 2.5 Workload Clusters on vSphere with the Tanzu CLI.

Note

This procedure describes how to create a new management cluster that uses Node IPAM directly. For information about how to update an existing management cluster to use Node IPAM, see Migrate an Existing Management Cluster from DHCP to Node IPAM.

You cannot deploy a new management cluster with Node IPAM directly from the installer interface; you must deploy it from a configuration file. You can create the configuration file by running the installer interface, clicking Review Configuration > Export Configuration, and then editing the generated configuration file as described below.

To deploy a new management cluster that uses in-cluster IPAM for its nodes:

  1. As a prerequisite, gather IP addresses of nameservers to use for the cluster’s control plane and worker nodes. This is required because cluster nodes will no longer receive nameservers via DHCP to resolve names in vCenter.

    Multiple Nameservers: For each nameserver setting you can specify a single nameserver, or two nameservers for redundancy. With two nameservers, a node VM normally uses the first nameserver listed in the setting, and only uses the second nameserver if the first one fails. One common setup is to specify multiple nameservers for all nodes, and set one as primary for worker nodes and the other as primary for control plane nodes.

  2. Edit the management cluster configuration file to include settings like the following, as described in the Node IPAM table in the Configuration File Variable Reference:

    MANAGEMENT_NODE_IPAM_IP_POOL_GATEWAY: "10.10.10.1"
    MANAGEMENT_NODE_IPAM_IP_POOL_ADDRESSES: "10.10.10.2-10.10.10.100,10.10.10.105"
    MANAGEMENT_NODE_IPAM_IP_POOL_SUBNET_PREFIX: "24"
    CONTROL_PLANE_NODE_NAMESERVERS: "10.10.10.10,10.10.10.11"
    WORKER_NODE_NAMESERVERS: "10.10.10.10,10.10.10.11"
    

    Where CONTROL_PLANE_NODE_NAMESERVERS and WORKER_NODE_NAMESERVERS are the addresses of the nameservers to use. For the NODE_NAMESERVERS settings, you can specify a single nameserver or two comma-separated nameservers.

    Dual-Stack: To configure a management cluster with dual-stack Node IPAM networking, include MANAGEMENT_NODE_IPAM_* settings as described in Configure Node IPAM with Dual-Stack Networking.

Configure Cluster Node MTU

To configure the maximum transmission unit (MTU) for management clusters and workload clusters on a vSphere Standard Switch, set the VSPHERE_MTU variable. Setting VSPHERE_MTU is applicable to both management clusters and workload clusters.

If not specified, the default vSphere node MTU is 1500. The maximum value is 9000. For information about MTUs, see About vSphere Networking in the vSphere 8 documentation.

Management Clusters on vSphere with Tanzu

On vSphere 7 and vSphere 8, vSphere with Tanzu provides a built-in Supervisor that serves as a management cluster and provides a better experience than a standalone management cluster. Deploying a Tanzu Kubernetes Grid management cluster to vSphere 7 or vSphere 8 when the Supervisor is not present is supported, but the preferred option is to enable vSphere with Tanzu and use the Supervisor if possible. Azure VMware Solution does not support a Supervisor Cluster, so you need to deploy a management cluster. For information, see vSphere with Tanzu Supervisor is a Management Cluster.

If vSphere with Tanzu is enabled, the installer interface states that you can use the TKG Service as the preferred way to run Kubernetes workloads, in which case you do not need a standalone management cluster. It presents a choice:

  • Configure vSphere with Tanzu opens the vSphere Client so you can configure your Supervisor as described in Configuring and Managing a Supervisor in the vSphere 8 documentation.

  • Deploy TKG Management Cluster allows you to continue deploying a standalone management cluster, for vSphere 7 or vSphere 8, and as required for and Azure VMware Solution.

Run the tanzu mc create Command

After you have created or updated the cluster configuration file, you can deploy a management cluster by running the tanzu mc create --file CONFIG-FILE command, where CONFIG-FILE is the name of the configuration file. If your configuration file is the default ~/.config/tanzu/tkg/cluster-config.yaml, you can omit the --file option. If you would like to review the Kubernetes manifest that the tanzu mc create command will apply you can optionally use the --dry-run flag to print the manifest without making changes. This invocation will still run the validation checks described below before generating the Kubernetes manifest.

Caution

The tanzu mc create command takes time to complete. While tanzu mc create is running, do not run additional invocations of tanzu mc create on the same bootstrap machine to deploy multiple management clusters, change context, or edit ~/.kube-tkg/config.

To deploy a management cluster, run the tanzu mc create command. For example:

tanzu mc create --file path/to/cluster-config-file.yaml

Multi-AZ: To run the management cluster or its workload clusters across multiple availability zones, either now or later, include the --az-file option with the vsphere-zones.yaml file described in Create FailureDomain and DeploymentZone Objects in Kubernetes:

tanzu mc create --file path/to/cluster-config-file.yaml --az-file path/to/vsphere-zones.yaml
  • After you deploy a management cluster to vSphere you must configure the IP addresses of its control plane nodes to be static, as described in Configure DHCP Reservations for the Control Plane in Creating and Managing TKG 2.5 Workload Clusters on vSphere with the Tanzu CLI.

Validation Checks

When you run tanzu mc create, the command performs several validation checks before deploying the management cluster. It verifies that the target vSphere infrastructure meets the following requirements:

  • The vSphere credentials that you provided are valid.
  • Nodes meet the minimum size requirements.
  • Base image template exists in vSphere and is valid for the specified Kubernetes version.
  • Required resources including the resource pool, datastores, and folder exist in vSphere.
  • Multi-AZ Validation: If you are deploying the management cluster defining FailureDomains and DeploymentZones resources as described in Create FailureDomain and DeploymentZone Objects in Kubernetes and then referencing them with the --az-file option in the tanzu mc create command, the cluster creation process by default performs the following additional checks:

    • All vSphere zones and regions referenced in VSphereFailureDomain, VSphereDeploymentZone, and related Kubernetes objects exist in vSphere, are accessible, and have host groups, VM groups, and tags that also exist.
    • Tag categories for the tags exist in vSphere and are logically consistent.
    • vSphere AZ/zone and region configurations are logically consistent.
    • MATCHING_LABELS configuration settings correspond to labels in the VSphereDeploymentZone objects.
    • Zones referenced in Kubernetes are at the same level in vSphere.
    • ResourcePools and Folders referenced in the VSphereDeploymentZone resources exist in vSphere.

    To prevent the cluster creation process from verifying that vSphere zones and regions specified in the configuration all exist, are consistent, and are defined at the same level, set SKIP_MULTI_AZ_VERIFY to "true" in your local environment:

     ```
     export SKIP_MULTI_AZ_VERIFY="true"
     ```
    

    You cannot set this variable in the cluster configuration file.

    A typical scenario for using SKIP_MULTI_AZ_VERIFY is when you are deploying a standalone management cluster that you will use to create workload clusters running across across multiple AZs in the future, but the vSphere resources for the workload cluster AZs have not yet been set up.

If any of these conditions are not met, the tanzu mc create command fails.

Monitoring Progress

When you run tanzu mc create, you can follow the progress of the deployment of the management cluster in the terminal. The first run of tanzu mc create takes longer than subsequent runs because it has to pull the required Docker images into the image store on your bootstrap machine. Subsequent runs do not require this step, so are faster.

If tanzu mc create fails before the management cluster deploys, you should clean up artifacts on your bootstrap machine before you re-run tanzu mc create. See the Troubleshooting Management Cluster Issues topic for details. If the machine on which you run tanzu mc create shuts down or restarts before the local operations finish, the deployment will fail.

If the deployment succeeds, you see a confirmation message in the terminal:

Management cluster created! You can now create your first workload cluster by running tanzu cluster create [name] -f [file]

What to Do Next

For information about what happened during the deployment of the management cluster, how to connect kubectl to the management cluster, how to create namespaces, and how to register the management cluster with Tanzu Mission Control, see Examine and Register a Newly-Deployed Standalone Management Cluster.

check-circle-line exclamation-circle-line close-line
Scroll to top icon