1. Documentation
×

Installing a private cluster on GCP

Prerequisites

Private clusters

You can deploy a private OKD cluster that does not expose external endpoints. Private clusters are accessible from only an internal network and are not visible to the internet.

By default, OKD is provisioned to use publicly-accessible DNS and endpoints. A private cluster sets the DNS, Ingress Controller, and API server to private when you deploy your cluster. This means that the cluster resources are only accessible from your internal network and are not visible to the internet.

If the cluster has any public subnets, load balancer services created by administrators might be publicly accessible. To ensure cluster security, verify that these services are explicitly annotated as private.

To deploy a private cluster, you must:

  • Use existing networking that meets your requirements. Your cluster resources might be shared between other clusters on the network.

  • Deploy from a machine that has access to:

    • The API services for the cloud to which you provision.

    • The hosts on the network that you provision.

    • The internet to obtain installation media.

You can use any machine that meets these access requirements and follows your company’s guidelines. For example, this machine can be a bastion host on your cloud network or a machine that has access to the network through a VPN.

Private clusters in GCP

To create a private cluster on Google Cloud Platform (GCP), you must provide an existing private VPC and subnets to host the cluster. The installation program must also be able to resolve the DNS records that the cluster requires. The installation program configures the Ingress Operator and API server for only internal traffic.

The cluster still requires access to internet to access the GCP APIs.

The following items are not required or created when you install a private cluster:

  • Public subnets

  • Public network load balancers, which support public ingress

  • A public DNS zone that matches the baseDomain for the cluster

The installation program does use the baseDomain that you specify to create a private DNS zone and the required records for the cluster. The cluster is configured so that the Operators do not create public records for the cluster and all cluster machines are placed in the private subnets that you specify.

Because it is not possible to limit access to external load balancers based on source tags, the private cluster uses only internal load balancers to allow access to internal instances.

The internal load balancer relies on instance groups rather than the target pools that the network load balancers use. The installation program creates instance groups for each zone, even if there is no instance in that group.

  • The cluster IP address is internal only.

  • One forwarding rule manages both the Kubernetes API and machine config server ports.

  • The backend service is comprised of each zone’s instance group and, while it exists, the bootstrap instance group.

  • The firewall uses a single rule that is based on only internal source ranges.

Limitations

No health check for the Machine config server, /healthz, runs because of a difference in load balancer functionality. Two internal load balancers cannot share a single IP address, but two network load balancers can share a single external IP address. Instead, the health of an instance is determined entirely by the /readyz check on port 6443.

About using a custom VPC

In OKD 4.12, you can deploy a cluster into an existing VPC in Google Cloud Platform (GCP). If you do, you must also use existing subnets within the VPC and routing rules.

By deploying OKD into an existing GCP VPC, you might be able to avoid limit constraints in new accounts or more easily abide by the operational constraints that your company’s guidelines set. This is a good option to use if you cannot obtain the infrastructure creation permissions that are required to create the VPC yourself.

Requirements for using your VPC

The installation program will no longer create the following components:

  • VPC

  • Subnets

  • Cloud router

  • Cloud NAT

  • NAT IP addresses

If you use a custom VPC, you must correctly configure it and its subnets for the installation program and the cluster to use. The installation program cannot subdivide network ranges for the cluster to use, set route tables for the subnets, or set VPC options like DHCP, so you must do so before you install the cluster.

Your VPC and subnets must meet the following characteristics:

  • The VPC must be in the same GCP project that you deploy the OKD cluster to.

  • To allow access to the internet from the control plane and compute machines, you must configure cloud NAT on the subnets to allow egress to it. These machines do not have a public address. Even if you do not require access to the internet, you must allow egress to the VPC network to obtain the installation program and images. Because multiple cloud NATs cannot be configured on the shared subnets, the installation program cannot configure it.

To ensure that the subnets that you provide are suitable, the installation program confirms the following data:

  • All the subnets that you specify exist and belong to the VPC that you specified.

  • The subnet CIDRs belong to the machine CIDR.

  • You must provide a subnet to deploy the cluster control plane and compute machines to. You can use the same subnet for both machine types.

If you destroy a cluster that uses an existing VPC, the VPC is not deleted.

Division of permissions

Starting with OKD 4.3, you do not need all of the permissions that are required for an installation program-provisioned infrastructure cluster to deploy a cluster. This change mimics the division of permissions that you might have at your company: some individuals can create different resources in your clouds than others. For example, you might be able to create application-specific items, like instances, buckets, and load balancers, but not networking-related components such as VPCs, subnets, or Ingress rules.

The GCP credentials that you use when you create your cluster do not need the networking permissions that are required to make VPCs and core networking components within the VPC, such as subnets, routing tables, internet gateways, NAT, and VPN. You still need permission to make the application resources that the machines within the cluster require, such as load balancers, security groups, storage, and nodes.

Isolation between clusters

If you deploy OKD to an existing network, the isolation of cluster services is preserved by firewall rules that reference the machines in your cluster by the cluster’s infrastructure ID. Only traffic within the cluster is allowed.

If you deploy multiple clusters to the same VPC, the following components might share access between clusters:

  • The API, which is globally available with an external publishing strategy or available throughout the network in an internal publishing strategy

  • Debugging tools, such as ports on VM instances that are open to the machine CIDR for SSH and ICMP access

Generating a key pair for cluster node SSH access

During an OKD installation, you can provide an SSH public key to the installation program. The key is passed to the Fedora CoreOS (FCOS) nodes through their Ignition config files and is used to authenticate SSH access to the nodes. The key is added to the ~/.ssh/authorized_keys list for the core user on each node, which enables password-less authentication.

After the key is passed to the nodes, you can use the key pair to SSH in to the FCOS nodes as the user core. To access the nodes through SSH, the private key identity must be managed by SSH for your local user.

If you want to SSH in to your cluster nodes to perform installation debugging or disaster recovery, you must provide the SSH public key during the installation process. The ./openshift-install gather command also requires the SSH public key to be in place on the cluster nodes.

Do not skip this procedure in production environments, where disaster recovery and debugging is required.

You must use a local key, not one that you configured with platform-specific approaches such as AWS key pairs.

On clusters running Fedora CoreOS (FCOS), the SSH keys specified in the Ignition config files are written to the /home/core/.ssh/authorized_keys.d/core file. However, the Machine Config Operator manages SSH keys in the /home/core/.ssh/authorized_keys file and configures sshd to ignore the /home/core/.ssh/authorized_keys.d/core file. As a result, newly provisioned OKD nodes are not accessible using SSH until the Machine Config Operator reconciles the machine configs with the authorized_keys file. After you can access the nodes using SSH, you can delete the /home/core/.ssh/authorized_keys.d/core file.
Procedure

  1. If you do not have an existing SSH key pair on your local machine to use for authentication onto your cluster nodes, create one. For example, on a computer that uses a Linux operating system, run the following command:

    $ ssh-keygen -t ed25519 -N '' -f <path>/<file_name> (1)
    1 Specify the path and file name, such as ~/.ssh/id_ed25519, of the new SSH key. If you have an existing key pair, ensure your public key is in the your ~/.ssh directory.

    If you plan to install an OKD cluster that uses FIPS validated or Modules In Process cryptographic libraries on the x86_64, ppc64le, and s390x architectures. do not create a key that uses the ed25519 algorithm. Instead, create a key that uses the rsa or ecdsa algorithm.

  2. View the public SSH key:

    $ cat <path>/<file_name>.pub

    For example, run the following to view the ~/.ssh/id_ed25519.pub public key:

    $ cat ~/.ssh/id_ed25519.pub

  3. Add the SSH private key identity to the SSH agent for your local user, if it has not already been added. SSH agent management of the key is required for password-less SSH authentication onto your cluster nodes, or if you want to use the ./openshift-install gather command.

    On some distributions, default SSH private key identities such as ~/.ssh/id_rsa and ~/.ssh/id_dsa are managed automatically.

    1. If the ssh-agent process is not already running for your local user, start it as a background task:

      $ eval "$(ssh-agent -s)"
      Example output
      Agent pid 31874

      If your cluster is in FIPS mode, only use FIPS-compliant algorithms to generate the SSH key. The key must be either RSA or ECDSA.

  4. Add your SSH private key to the ssh-agent:

    $ ssh-add <path>/<file_name> (1)
    1 Specify the path and file name for your SSH private key, such as ~/.ssh/id_ed25519
    Example output
    Identity added: /home/<you>/<path>/<file_name> (<computer_name>)
Next steps

  • When you install OKD, provide the SSH public key to the installation program.

Obtaining the installation program

Before you install OKD, download the installation file on the host you are using for installation.

Prerequisites

  • You have a computer that runs Linux or macOS, with 500 MB of local disk space.

Procedure

  1. Download installer from https://github.com/openshift/okd/releases

    The installation program creates several files on the computer that you use to install your cluster. You must keep the installation program and the files that the installation program creates after you finish installing the cluster. Both files are required to delete the cluster.

    Deleting the files created by the installation program does not remove your cluster, even if the cluster failed during installation. To remove your cluster, complete the OKD uninstallation procedures for your specific cloud provider.

  2. Extract the installation program. For example, on a computer that uses a Linux operating system, run the following command:

    $ tar -xvf openshift-install-linux.tar.gz

  3. Download your installation pull secret from the Red Hat OpenShift Cluster Manager. This pull secret allows you to authenticate with the services that are provided by the included authorities, including Quay.io, which serves the container images for OKD components.

    Using a pull secret from the Red Hat OpenShift Cluster Manager is not required. You can use a pull secret for another private registry. Or, if you do not need the cluster to pull images from a private registry, you can use {"auths":{"fake":{"auth":"aWQ6cGFzcwo="}}} as the pull secret when prompted during the installation.

    If you do not use the pull secret from the Red Hat OpenShift Cluster Manager:

    • Red Hat Operators are not available.

    • The Telemetry and Insights operators do not send data to Red Hat.

    • Content from the Red Hat Container Catalog registry, such as image streams and Operators, are not available.

Manually creating the installation configuration file

Installing the cluster requires that you manually create the installation configuration file.

Prerequisites

  • You have an SSH public key on your local machine to provide to the installation program. The key will be used for SSH authentication onto your cluster nodes for debugging and disaster recovery.

  • You have obtained the OKD installation program and the pull secret for your cluster.

Procedure

  1. Create an installation directory to store your required installation assets in:

    $ mkdir <installation_directory>

    You must create a directory. Some installation assets, like bootstrap X.509 certificates have short expiration intervals, so you must not reuse an installation directory. If you want to reuse individual files from another cluster installation, you can copy them into your directory. However, the file names for the installation assets might change between releases. Use caution when copying installation files from an earlier OKD version.

  2. Customize the sample install-config.yaml file template that is provided and save it in the <installation_directory>.

    You must name this configuration file install-config.yaml.

  3. Back up the install-config.yaml file so that you can use it to install multiple clusters.

    The install-config.yaml file is consumed during the next step of the installation process. You must back it up now.

Installation configuration parameters

Before you deploy an OKD cluster, you provide parameter values to describe your account on the cloud platform that hosts your cluster and optionally customize your cluster’s platform. When you create the install-config.yaml installation configuration file, you provide values for the required parameters through the command line. If you customize your cluster, you can modify the install-config.yaml file to provide more details about the platform.

After installation, you cannot modify these parameters in the install-config.yaml file.

Required configuration parameters

Required installation configuration parameters are described in the following table:

Table 1. Required parameters
Parameter Description Values

apiVersion

The API version for the install-config.yaml content. The current version is v1. The installation program may also support older API versions.

String

baseDomain

The base domain of your cloud provider. The base domain is used to create routes to your OKD cluster components. The full DNS name for your cluster is a combination of the baseDomain and metadata.name parameter values that uses the <metadata.name>.<baseDomain> format.

A fully-qualified domain or subdomain name, such as example.com.

metadata

Kubernetes resource ObjectMeta, from which only the name parameter is consumed.

Object

metadata.name

The name of the cluster. DNS records for the cluster are all subdomains of {{.metadata.name}}.{{.baseDomain}}.

String of lowercase letters, hyphens (-), and periods (.), such as dev.

platform

The configuration for the specific platform upon which to perform the installation: alibabacloud, aws, baremetal, azure, gcp, ibmcloud, nutanix, openstack, ovirt, vsphere, or {}. For additional information about platform.<platform> parameters, consult the table for your specific platform that follows.

Object

Network configuration parameters

You can customize your installation configuration based on the requirements of your existing network infrastructure. For example, you can expand the IP address block for the cluster network or provide different IP address blocks than the defaults.

Only IPv4 addresses are supported.

Globalnet is not supported with Red Hat OpenShift Data Foundation disaster recovery solutions. For regional disaster recovery scenarios, ensure that you use a nonoverlapping range of private IP addresses for the cluster and service networks in each cluster.
Table 2. Network parameters
Parameter Description Values

networking

The configuration for the cluster network.

Object

You cannot modify parameters specified by the networking object after installation.

networking.networkType

The Red Hat OpenShift Networking network plugin to install.

Either OpenShiftSDN or OVNKubernetes. The default value is OVNKubernetes.

networking.clusterNetwork

The IP address blocks for pods.

The default value is 10.128.0.0/14 with a host prefix of /23.

If you specify multiple IP address blocks, the blocks must not overlap.

An array of objects. For example:

networking:
  clusterNetwork:
  - cidr: 10.128.0.0/14
    hostPrefix: 23

networking.clusterNetwork.cidr

Required if you use networking.clusterNetwork. An IP address block.

An IPv4 network.

An IP address block in Classless Inter-Domain Routing (CIDR) notation. The prefix length for an IPv4 block is between 0 and 32.

networking.clusterNetwork.hostPrefix

The subnet prefix length to assign to each individual node. For example, if hostPrefix is set to 23 then each node is assigned a /23 subnet out of the given cidr. A hostPrefix value of 23 provides 510 (2^(32 - 23) - 2) pod IP addresses.

A subnet prefix.

The default value is 23.

networking.serviceNetwork

The IP address block for services. The default value is 172.30.0.0/16.

The OpenShift SDN and OVN-Kubernetes network plugins support only a single IP address block for the service network.

An array with an IP address block in CIDR format. For example:

networking:
  serviceNetwork:
   - 172.30.0.0/16

networking.machineNetwork

The IP address blocks for machines.

If you specify multiple IP address blocks, the blocks must not overlap.

An array of objects. For example:

networking:
  machineNetwork:
  - cidr: 10.0.0.0/16

networking.machineNetwork.cidr

Required if you use networking.machineNetwork. An IP address block. The default value is 10.0.0.0/16 for all platforms other than libvirt. For libvirt, the default value is 192.168.126.0/24.

An IP network block in CIDR notation.

For example, 10.0.0.0/16.

Set the networking.machineNetwork to match the CIDR that the preferred NIC resides in.

Optional configuration parameters

Optional installation configuration parameters are described in the following table:

Table 3. Optional parameters
Parameter Description Values

additionalTrustBundle

A PEM-encoded X.509 certificate bundle that is added to the nodes' trusted certificate store. This trust bundle may also be used when a proxy has been configured.

String

capabilities

Controls the installation of optional core cluster components. You can reduce the footprint of your OKD cluster by disabling optional components. For more information, see the "Cluster capabilities" page in Installing.

String array

capabilities.baselineCapabilitySet

Selects an initial set of optional capabilities to enable. Valid values are None, v4.11, v4.12 and vCurrent. The default value is vCurrent.

String

capabilities.additionalEnabledCapabilities

Extends the set of optional capabilities beyond what you specify in baselineCapabilitySet. You may specify multiple capabilities in this parameter.

String array

compute

The configuration for the machines that comprise the compute nodes.

Array of MachinePool objects.

compute.architecture

Determines the instruction set architecture of the machines in the pool. Currently, clusters with varied architectures are not supported. All pools must specify the same architecture. Valid values are amd64 (the default).

String

compute.hyperthreading

Whether to enable or disable simultaneous multithreading, or hyperthreading, on compute machines. By default, simultaneous multithreading is enabled to increase the performance of your machines' cores.

If you disable simultaneous multithreading, ensure that your capacity planning accounts for the dramatically decreased machine performance.

Enabled or Disabled

compute.name

Required if you use compute. The name of the machine pool.

worker

compute.platform

Required if you use compute. Use this parameter to specify the cloud provider to host the worker machines. This parameter value must match the controlPlane.platform parameter value.

alibabacloud, aws, azure, gcp, ibmcloud, nutanix, openstack, ovirt, vsphere, or {}

compute.replicas

The number of compute machines, which are also known as worker machines, to provision.

A positive integer greater than or equal to 2. The default value is 3.

featureSet

Enables the cluster for a feature set. A feature set is a collection of OKD features that are not enabled by default. For more information about enabling a feature set during installation, see "Enabling features using feature gates".

String. The name of the feature set to enable, such as TechPreviewNoUpgrade.

controlPlane

The configuration for the machines that comprise the control plane.

Array of MachinePool objects.

controlPlane.architecture

Determines the instruction set architecture of the machines in the pool. Currently, clusters with varied architectures are not supported. All pools must specify the same architecture. Valid values are amd64.

String

controlPlane.hyperthreading

Whether to enable or disable simultaneous multithreading, or hyperthreading, on control plane machines. By default, simultaneous multithreading is enabled to increase the performance of your machines' cores.

If you disable simultaneous multithreading, ensure that your capacity planning accounts for the dramatically decreased machine performance.

Enabled or Disabled

controlPlane.name

Required if you use controlPlane. The name of the machine pool.

master

controlPlane.platform

Required if you use controlPlane. Use this parameter to specify the cloud provider that hosts the control plane machines. This parameter value must match the compute.platform parameter value.

alibabacloud, aws, azure, gcp, ibmcloud, nutanix, openstack, ovirt, vsphere, or {}

controlPlane.replicas

The number of control plane machines to provision.

The only supported value is 3, which is the default value.

credentialsMode

The Cloud Credential Operator (CCO) mode. If no mode is specified, the CCO dynamically tries to determine the capabilities of the provided credentials, with a preference for mint mode on the platforms where multiple modes are supported. If you are installing on GCP into a shared virtual private cloud (VPC), credentialsMode must be set to Passthrough.

Not all CCO modes are supported for all cloud providers. For more information about CCO modes, see the Cloud Credential Operator entry in the Cluster Operators reference content.

If your AWS account has service control policies (SCP) enabled, you must configure the credentialsMode parameter to Mint, Passthrough or Manual.

Mint, Passthrough, Manual or an empty string ("").

imageContentSources

Sources and repositories for the release-image content.

Array of objects. Includes a source and, optionally, mirrors, as described in the following rows of this table.

imageContentSources.source

Required if you use imageContentSources. Specify the repository that users refer to, for example, in image pull specifications.

String

imageContentSources.mirrors

Specify one or more repositories that may also contain the same images.

Array of strings

publish

How to publish or expose the user-facing endpoints of your cluster, such as the Kubernetes API, OpenShift routes.

Internal or External. To deploy a private cluster, which cannot be accessed from the internet, set publish to Internal. The default value is External.

sshKey

The SSH key to authenticate access to your cluster machines.

For production OKD clusters on which you want to perform installation debugging or disaster recovery, specify an SSH key that your ssh-agent process uses.

For example, sshKey: ssh-ed25519 AAAA...

Additional Google Cloud Platform (GCP) configuration parameters

Additional GCP configuration parameters are described in the following table:

Table 4. Additional GCP parameters
Parameter Description Values

platform.gcp.network

The name of the existing Virtual Private Cloud (VPC) where you want to deploy your cluster. If you want to deploy your cluster into a shared VPC, you must set platform.gcp.networkProjectID with the name of the GCP project that contains the shared VPC.

String.

platform.gcp.networkProjectID

Optional. The name of the GCP project that contains the shared VPC where you want to deploy your cluster.

String.

platform.gcp.projectID

The name of the GCP project where the installation program installs the cluster.

String.

platform.gcp.region

The name of the GCP region that hosts your cluster.

Any valid region name, such as us-central1.

platform.gcp.controlPlaneSubnet

The name of the existing subnet where you want to deploy your control plane machines.

The subnet name.

platform.gcp.computeSubnet

The name of the existing subnet where you want to deploy your compute machines.

The subnet name.

platform.gcp.createFirewallRules

Optional. Set this value to Disabled if you want to create and manage your firewall rules using network tags. By default, the cluster will automatically create and manage the firewall rules that are required for cluster communication. Your service account must have roles/compute.networkAdmin and roles/compute.securityAdmin privileges in the host project to perform these tasks automatically. If your service account does not have the roles/dns.admin privilege in the host project, it must have the dns.networks.bindPrivateDNSZone permission.

Enabled or Disabled. The default value is Enabled.

platform.gcp.publicDNSZone.project

Optional. The name of the project that contains the public DNS zone. If you set this value, your service account must have the roles/dns.admin privilege in the specified project. If you do not set this value, it defaults to gcp.projectId.

The name of the project that contains the public DNS zone.

platform.gcp.publicDNSZone.id

Optional. The ID or name of an existing public DNS zone. The public DNS zone domain must match the baseDomain parameter. If you do not set this value, the installation program will use a public DNS zone in the service project.

The public DNS zone name.

platform.gcp.privateDNSZone.project

Optional. The name of the project that contains the private DNS zone. If you set this value, your service account must have the roles/dns.admin privilege in the host project. If you do not set this value, it defaults to gcp.projectId.

The name of the project that contains the private DNS zone.

platform.gcp.privateDNSZone.id

Optional. The ID or name of an existing private DNS zone. If you do not set this value, the installation program will create a private DNS zone in the service project.

The private DNS zone name.

platform.gcp.licenses

A list of license URLs that must be applied to the compute images.

The licenses parameter is a deprecated field and nested virtualization is enabled by default. It is not recommended to use this field.

Any license available with the license API, such as the license to enable nested virtualization. You cannot use this parameter with a mechanism that generates pre-built images. Using a license URL forces the installation program to copy the source image before use.

platform.gcp.defaultMachinePlatform.zones

The availability zones where the installation program creates machines.

A list of valid GCP availability zones, such as us-central1-a, in a YAML sequence.

platform.gcp.defaultMachinePlatform.osDisk.diskSizeGB

The size of the disk in gigabytes (GB).

Any size between 16 GB and 65536 GB.

platform.gcp.defaultMachinePlatform.osDisk.diskType

The GCP disk type.

Either the default pd-ssd or the pd-standard disk type. The control plane nodes must be the pd-ssd disk type. Compute nodes can be either type.

platform.gcp.defaultMachinePlatform.osImage.project

Optional. By default, the installation program downloads and installs the FCOS image that is used to boot control plane and compute machines. You can override the default behavior by specifying the location of a custom FCOS image for the installation program to use for both types of machines.

String. The name of GCP project where the image is located.

platform.gcp.defaultMachinePlatform.osImage.name

The name of the custom FCOS image for the installation program to use to boot control plane and compute machines. If you use platform.gcp.defaultMachinePlatform.osImage.project, this field is required.

String. The name of the RHCOS image.

platform.gcp.defaultMachinePlatform.tags

Optional. Additional network tags to add to the control plane and compute machines.

One or more strings, for example network-tag1.

platform.gcp.defaultMachinePlatform.type

The GCP machine type for control plane and compute machines.

The GCP machine type, for example n1-standard-4.

platform.gcp.defaultMachinePlatform.osDisk.encryptionKey.kmsKey.name

The name of the customer managed encryption key to be used for machine disk encryption.

The encryption key name.

platform.gcp.defaultMachinePlatform.osDisk.encryptionKey.kmsKey.keyRing

The name of the Key Management Service (KMS) key ring to which the KMS key belongs.

The KMS key ring name.

platform.gcp.defaultMachinePlatform.osDisk.encryptionKey.kmsKey.location

The GCP location in which the KMS key ring exists.

The GCP location.

platform.gcp.defaultMachinePlatform.osDisk.encryptionKey.kmsKey.projectID

The ID of the project in which the KMS key ring exists. This value defaults to the value of the platform.gcp.projectID parameter if it is not set.

The GCP project ID.

platform.gcp.defaultMachinePlatform.osDisk.encryptionKey.kmsKeyServiceAccount

The GCP service account used for the encryption request for control plane and compute machines. If absent, the Compute Engine default service account is used. For more information about GCP service accounts, see Google’s documentation on service accounts.

The GCP service account email, for example <service_account_name>@<project_id>.iam.gserviceaccount.com.

controlPlane.platform.gcp.osDisk.encryptionKey.kmsKey.name

The name of the customer managed encryption key to be used for control plane machine disk encryption.

The encryption key name.

controlPlane.platform.gcp.osDisk.encryptionKey.kmsKey.keyRing

For control plane machines, the name of the KMS key ring to which the KMS key belongs.

The KMS key ring name.

controlPlane.platform.gcp.osDisk.encryptionKey.kmsKey.location

For control plane machines, the GCP location in which the key ring exists. For more information about KMS locations, see Google’s documentation on Cloud KMS locations.

The GCP location for the key ring.

controlPlane.platform.gcp.osDisk.encryptionKey.kmsKey.projectID

For control plane machines, the ID of the project in which the KMS key ring exists. This value defaults to the VM project ID if not set.

The GCP project ID.

controlPlane.platform.gcp.osDisk.encryptionKey.kmsKeyServiceAccount

The GCP service account used for the encryption request for control plane machines. If absent, the Compute Engine default service account is used. For more information about GCP service accounts, see Google’s documentation on service accounts.

The GCP service account email, for example <service_account_name>@<project_id>.iam.gserviceaccount.com.

controlPlane.platform.gcp.osDisk.diskSizeGB

The size of the disk in gigabytes (GB). This value applies to control plane machines.

Any integer between 16 and 65536.

controlPlane.platform.gcp.osDisk.diskType

The GCP disk type for control plane machines.

Control plane machines must use the pd-ssd disk type, which is the default.

controlPlane.platform.gcp.osImage.project

Optional. By default, the installation program downloads and installs the Fedora CoreOS (FCOS) image that is used to boot control plane machines. You can override the default behavior by specifying the location of a custom FCOS image for the installation program to use for control plane machines only.

String. The name of GCP project where the image is located.

controlPlane.platform.gcp.osImage.name

The name of the custom FCOS image for the installation program to use to boot control plane machines. If you use controlPlane.platform.gcp.osImage.project, this field is required.

String. The name of the FCOS image.

controlPlane.platform.gcp.tags

Optional. Additional network tags to add to the control plane machines. If set, this parameter overrides the platform.gcp.defaultMachinePlatform.tags parameter for control plane machines.

One or more strings, for example control-plane-tag1.

controlPlane.platform.gcp.type

The GCP machine type for control plane machines. If set, this parameter overrides the platform.gcp.defaultMachinePlatform.type parameter.

The GCP machine type, for example n1-standard-4.

controlPlane.platform.gcp.zones

The availability zones where the installation program creates control plane machines.

A list of valid GCP availability zones, such as us-central1-a, in a YAML sequence.

compute.platform.gcp.osDisk.encryptionKey.kmsKey.name

The name of the customer managed encryption key to be used for compute machine disk encryption.

The encryption key name.

compute.platform.gcp.osDisk.encryptionKey.kmsKey.keyRing

For compute machines, the name of the KMS key ring to which the KMS key belongs.

The KMS key ring name.

compute.platform.gcp.osDisk.encryptionKey.kmsKey.location

For compute machines, the GCP location in which the key ring exists. For more information about KMS locations, see Google’s documentation on Cloud KMS locations.

The GCP location for the key ring.

compute.platform.gcp.osDisk.encryptionKey.kmsKey.projectID

For compute machines, the ID of the project in which the KMS key ring exists. This value defaults to the VM project ID if not set.

The GCP project ID.

compute.platform.gcp.osDisk.encryptionKey.kmsKeyServiceAccount

The GCP service account used for the encryption request for compute machines. If this value is not set, the Compute Engine default service account is used. For more information about GCP service accounts, see Google’s documentation on service accounts.

The GCP service account email, for example <service_account_name>@<project_id>.iam.gserviceaccount.com.

compute.platform.gcp.osDisk.diskSizeGB

The size of the disk in gigabytes (GB). This value applies to compute machines.

Any integer between 16 and 65536.

compute.platform.gcp.osDisk.diskType

The GCP disk type for compute machines.

Either the default pd-ssd or the pd-standard disk type.

compute.platform.gcp.osImage.project

Optional. By default, the installation program downloads and installs the FCOS image that is used to boot compute machines. You can override the default behavior by specifying the location of a custom FCOS image for the installation program to use for compute machines only.

String. The name of GCP project where the image is located.

compute.platform.gcp.osImage.name

The name of the custom FCOS image for the installation program to use to boot compute machines. If you use compute.platform.gcp.osImage.project, this field is required.

String. The name of the FCOS image.

compute.platform.gcp.tags

Optional. Additional network tags to add to the compute machines. If set, this parameter overrides the platform.gcp.defaultMachinePlatform.tags parameter for compute machines.

One or more strings, for example compute-network-tag1.

compute.platform.gcp.type

The GCP machine type for compute machines. If set, this parameter overrides the platform.gcp.defaultMachinePlatform.type parameter.

The GCP machine type, for example n1-standard-4.

compute.platform.gcp.zones

The availability zones where the installation program creates compute machines.

A list of valid GCP availability zones, such as us-central1-a, in a YAML sequence.

Minimum resource requirements for cluster installation

Each cluster machine must meet the following minimum requirements:

Table 5. Minimum resource requirements
Machine Operating System vCPU [1] Virtual RAM Storage Input/Output Per Second (IOPS)[2]

Bootstrap

FCOS

4

16 GB

100 GB

300

Control plane

FCOS

4

16 GB

100 GB

300

Compute

FCOS

2

8 GB

100 GB

300

  1. One vCPU is equivalent to one physical core when simultaneous multithreading (SMT), or hyperthreading, is not enabled. When enabled, use the following formula to calculate the corresponding ratio: (threads per core × cores) × sockets = vCPUs.

  2. OKD and Kubernetes are sensitive to disk performance, and faster storage is recommended, particularly for etcd on the control plane nodes which require a 10 ms p99 fsync duration. Note that on many cloud platforms, storage size and IOPS scale together, so you might need to over-allocate storage volume to obtain sufficient performance.

  3. As with all user-provisioned installations, if you choose to use Fedora compute machines in your cluster, you take responsibility for all operating system life cycle management and maintenance, including performing system updates, applying patches, and completing all other required tasks. Use of Fedora 7 compute machines is deprecated and has been removed in OKD 4.10 and later.

If an instance type for your platform meets the minimum requirements for cluster machines, it is supported to use in OKD.

Additional resources

Tested instance types for GCP

The following Google Cloud Platform instance types have been tested with OKD.

Machine series

  • A2

  • A3

  • C2

  • C2D

  • C3

  • C3D

  • C4

  • E2

  • M1

  • N1

  • N2

  • N2D

  • N4

  • Tau T2D

Using custom machine types

Using a custom machine type to install a OKD cluster is supported.

Consider the following when using a custom machine type:

  • Similar to predefined instance types, custom machine types must meet the minimum resource requirements for control plane and compute machines. For more information, see "Minimum resource requirements for cluster installation".

  • The name of the custom machine type must adhere to the following syntax:

    custom-<number_of_cpus>-<amount_of_memory_in_mb>

    For example, custom-6-20480.

As part of the installation process, you specify the custom machine type in the install-config.yaml file.

Sample install-config.yaml file with a custom machine type
compute:
- architecture: amd64
  hyperthreading: Enabled
  name: worker
  platform:
    gcp:
      type: custom-6-20480
  replicas: 2
controlPlane:
  architecture: amd64
  hyperthreading: Enabled
  name: master
  platform:
    gcp:
      type: custom-6-20480
  replicas: 3

Sample customized install-config.yaml file for GCP

You can customize the install-config.yaml file to specify more details about your OKD cluster’s platform or modify the values of the required parameters.

This sample YAML file is provided for reference only. You must obtain your install-config.yaml file by using the installation program and modify it.

 
apiVersion: v1
baseDomain: example.com (1)
controlPlane:  (2) (3)
  hyperthreading: Enabled (4)
  name: master
  platform:
    gcp:
      type: n2-standard-4
      zones:
      - us-central1-a
      - us-central1-c
      osDisk:
        diskType: pd-ssd
        diskSizeGB: 1024
        encryptionKey: (5)
          kmsKey:
            name: worker-key
            keyRing: test-machine-keys
            location: global
            projectID: project-id
      tags: (6)
      - control-plane-tag1
      - control-plane-tag2
      osImage: (7)
        project: example-project-name
        name: example-image-name
  replicas: 3
compute:  (2) (3)
- hyperthreading: Enabled (4)
  name: worker
  platform:
    gcp:
      type: n2-standard-4
      zones:
      - us-central1-a
      - us-central1-c
      osDisk:
        diskType: pd-standard
        diskSizeGB: 128
        encryptionKey: (5)
          kmsKey:
            name: worker-key
            keyRing: test-machine-keys
            location: global
            projectID: project-id
      tags: (6)
      - compute-tag1
      - compute-tag2
      osImage: (7)
          project: example-project-name
          name: example-image-name
  replicas: 3
metadata:
  name: test-cluster (1)
networking:
  clusterNetwork:
  - cidr: 10.128.0.0/14
    hostPrefix: 23
  machineNetwork:
  - cidr: 10.0.0.0/16
  networkType: OVNKubernetes (8)
  serviceNetwork:
  - 172.30.0.0/16
platform:
  gcp:
    projectID: openshift-production (1)
    region: us-central1 (1)
    defaultMachinePlatform:
      tags: (6)
      - global-tag1
      - global-tag2
      osImage: (7)
        project: example-project-name
        name: example-image-name
    network: existing_vpc (9)
    controlPlaneSubnet: control_plane_subnet (10)
    computeSubnet: compute_subnet (11)
pullSecret: '{"auths": ...}' (1)
sshKey: ssh-ed25519 AAAA... (12)
publish: Internal (13)
1 Required. The installation program prompts you for this value.
2 If you do not provide these parameters and values, the installation program provides the default value.
3 The controlPlane section is a single mapping, but the compute section is a sequence of mappings. To meet the requirements of the different data structures, the first line of the compute section must begin with a hyphen, -, and the first line of the controlPlane section must not. Only one control plane pool is used.
4 Whether to enable or disable simultaneous multithreading, or hyperthreading. By default, simultaneous multithreading is enabled to increase the performance of your machines' cores. You can disable it by setting the parameter value to Disabled. If you disable simultaneous multithreading in some cluster machines, you must disable it in all cluster machines.

If you disable simultaneous multithreading, ensure that your capacity planning accounts for the dramatically decreased machine performance. Use larger machine types, such as n1-standard-8, for your machines if you disable simultaneous multithreading.
5 Optional: The custom encryption key section to encrypt both virtual machines and persistent volumes. Your default compute service account must have the permissions granted to use your KMS key and have the correct IAM role assigned. The default service account name follows the service-<project_number>@compute-system.iam.gserviceaccount.com pattern. For more information about granting the correct permissions for your service account, see "Machine management" → "Creating compute machine sets" → "Creating a compute machine set on GCP".
6 Optional: A set of network tags to apply to the control plane or compute machine sets. The platform.gcp.defaultMachinePlatform.tags parameter will apply to both control plane and compute machines. If the compute.platform.gcp.tags or controlPlane.platform.gcp.tags parameters are set, they override the platform.gcp.defaultMachinePlatform.tags parameter.
7 Optional: A custom Fedora CoreOS (FCOS) image for the installation program to use to boot control plane and compute machines. The project and name parameters under platform.gcp.defaultMachinePlatform.osImage apply to both control plane and compute machines. If the project and name parameters under controlPlane.platform.gcp.osImage or compute.platform.gcp.osImage are set, they override the platform.gcp.defaultMachinePlatform.osImage parameters.
8 The cluster network plugin to install. The supported values are OVNKubernetes and OpenShiftSDN. The default value is OVNKubernetes.
9 Specify the name of an existing VPC.
10 Specify the name of the existing subnet to deploy the control plane machines to. The subnet must belong to the VPC that you specified.
11 Specify the name of the existing subnet to deploy the compute machines to. The subnet must belong to the VPC that you specified.
12 You can optionally provide the sshKey value that you use to access the machines in your cluster.

For production OKD clusters on which you want to perform installation debugging or disaster recovery, specify an SSH key that your ssh-agent process uses.
13 How to publish the user-facing endpoints of your cluster. Set publish to Internal to deploy a private cluster, which cannot be accessed from the internet. The default value is External.
Additional resources

Create an Ingress Controller with global access on GCP

You can create an Ingress Controller that has global access to a Google Cloud Platform (GCP) cluster. Global access is only available to Ingress Controllers using internal load balancers.

Prerequisites

  • You created the install-config.yaml and complete any modifications to it.

Procedure

Create an Ingress Controller with global access on a new GCP cluster.

  1. Change to the directory that contains the installation program and create a manifest file:

    $ ./openshift-install create manifests --dir <installation_directory> (1)
    1 For <installation_directory>, specify the name of the directory that contains the install-config.yaml file for your cluster.

  2. Create a file that is named cluster-ingress-default-ingresscontroller.yaml in the <installation_directory>/manifests/ directory:

    $ touch <installation_directory>/manifests/cluster-ingress-default-ingresscontroller.yaml (1)
    1 For <installation_directory>, specify the directory name that contains the manifests/ directory for your cluster.

    After creating the file, several network configuration files are in the manifests/ directory, as shown:

    $ ls <installation_directory>/manifests/cluster-ingress-default-ingresscontroller.yaml
    Example output
    cluster-ingress-default-ingresscontroller.yaml

  3. Open the cluster-ingress-default-ingresscontroller.yaml file in an editor and enter a custom resource (CR) that describes the Operator configuration you want:

    Sample clientAccess configuration to Global
      apiVersion: operator.openshift.io/v1
  kind: IngressController
  metadata:
    name: default
    namespace: openshift-ingress-operator
  spec:
    endpointPublishingStrategy:
      loadBalancer:
        providerParameters:
          gcp:
            clientAccess: Global (1)
          type: GCP
        scope: Internal          (2)
      type: LoadBalancerService
    1 Set gcp.clientAccess to Global.
    2 Global access is only available to Ingress Controllers using internal load balancers.

Configuring the cluster-wide proxy during installation

Production environments can deny direct access to the internet and instead have an HTTP or HTTPS proxy available. You can configure a new OKD cluster to use a proxy by configuring the proxy settings in the install-config.yaml file.

Prerequisites

  • You have an existing install-config.yaml file.

  • You reviewed the sites that your cluster requires access to and determined whether any of them need to bypass the proxy. By default, all cluster egress traffic is proxied, including calls to hosting cloud provider APIs. You added sites to the Proxy object’s spec.noProxy field to bypass the proxy if necessary.

    The Proxy object status.noProxy field is populated with the values of the networking.machineNetwork[].cidr, networking.clusterNetwork[].cidr, and networking.serviceNetwork[] fields from your installation configuration.

    For installations on Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure, and OpenStack, the Proxy object status.noProxy field is also populated with the instance metadata endpoint (169.254.169.254).
Procedure

  1. Edit your install-config.yaml file and add the proxy settings. For example:

    apiVersion: v1
baseDomain: my.domain.com
proxy:
  httpProxy: http://<username>:<pswd>@<ip>:<port> (1)
  httpsProxy: https://<username>:<pswd>@<ip>:<port> (2)
  noProxy: example.com (3)
additionalTrustBundle: | (4)
    -----BEGIN CERTIFICATE-----
    <MY_TRUSTED_CA_CERT>
    -----END CERTIFICATE-----
additionalTrustBundlePolicy: <policy_to_add_additionalTrustBundle> (5)
    1 A proxy URL to use for creating HTTP connections outside the cluster. The URL scheme must be http.
    2 A proxy URL to use for creating HTTPS connections outside the cluster.
    3 A comma-separated list of destination domain names, IP addresses, or other network CIDRs to exclude from proxying. Preface a domain with . to match subdomains only. For example, .y.com matches x.y.com, but not y.com. Use * to bypass the proxy for all destinations.
    4 If provided, the installation program generates a config map that is named user-ca-bundle in the openshift-config namespace that contains one or more additional CA certificates that are required for proxying HTTPS connections. The Cluster Network Operator then creates a trusted-ca-bundle config map that merges these contents with the Fedora CoreOS (FCOS) trust bundle, and this config map is referenced in the trustedCA field of the Proxy object. The additionalTrustBundle field is required unless the proxy’s identity certificate is signed by an authority from the FCOS trust bundle.
    5 Optional: The policy to determine the configuration of the Proxy object to reference the user-ca-bundle config map in the trustedCA field. The allowed values are Proxyonly and Always. Use Proxyonly to reference the user-ca-bundle config map only when http/https proxy is configured. Use Always to always reference the user-ca-bundle config map. The default value is Proxyonly.

    The installation program does not support the proxy readinessEndpoints field.

    If the installer times out, restart and then complete the deployment by using the wait-for command of the installer. For example:

    $ ./openshift-install wait-for install-complete --log-level debug

  2. Save the file and reference it when installing OKD.

The installation program creates a cluster-wide proxy that is named cluster that uses the proxy settings in the provided install-config.yaml file. If no proxy settings are provided, a cluster Proxy object is still created, but it will have a nil spec.

Only the Proxy object named cluster is supported, and no additional proxies can be created.

Deploying the cluster

You can install OKD on a compatible cloud platform.

You can run the create cluster command of the installation program only once, during initial installation.
Prerequisites

  • Configure an account with the cloud platform that hosts your cluster.

  • Obtain the OKD installation program and the pull secret for your cluster.

  • Verify the cloud provider account on your host has the correct permissions to deploy the cluster. An account with incorrect permissions causes the installation process to fail with an error message that displays the missing permissions.

Procedure

  1. Remove any existing GCP credentials that do not use the service account key for the GCP account that you configured for your cluster and that are stored in the following locations:

    • The GOOGLE_CREDENTIALS, GOOGLE_CLOUD_KEYFILE_JSON, or GCLOUD_KEYFILE_JSON environment variables

    • The ~/.gcp/osServiceAccount.json file

    • The gcloud cli default credentials

  2. Change to the directory that contains the installation program and initialize the cluster deployment:

    $ ./openshift-install create cluster --dir <installation_directory> \ (1)
    --log-level=info (2)
    1 For <installation_directory>, specify the location of your customized ./install-config.yaml file.
    2 To view different installation details, specify warn, debug, or error instead of info.

    If the cloud provider account that you configured on your host does not have sufficient permissions to deploy the cluster, the installation process stops, and the missing permissions are displayed.

  3. Optional: You can reduce the number of permissions for the service account that you used to install the cluster.

    • If you assigned the Owner role to your service account, you can remove that role and replace it with the Viewer role.

    • If you included the Service Account Key Admin role, you can remove it.

Verification

When the cluster deployment completes successfully:

  • The terminal displays directions for accessing your cluster, including a link to the web console and credentials for the kubeadmin user.

  • Credential information also outputs to <installation_directory>/.openshift_install.log.

Do not delete the installation program or the files that the installation program creates. Both are required to delete the cluster.
Example output
...
INFO Install complete!
INFO To access the cluster as the system:admin user when using 'oc', run 'export KUBECONFIG=/home/myuser/install_dir/auth/kubeconfig'
INFO Access the OpenShift web-console here: https://console-openshift-console.apps.mycluster.example.com
INFO Login to the console with user: "kubeadmin", and password: "password"
INFO Time elapsed: 36m22s

  • The Ignition config files that the installation program generates contain certificates that expire after 24 hours, which are then renewed at that time. If the cluster is shut down before renewing the certificates and the cluster is later restarted after the 24 hours have elapsed, the cluster automatically recovers the expired certificates. The exception is that you must manually approve the pending node-bootstrapper certificate signing requests (CSRs) to recover kubelet certificates. See the documentation for Recovering from expired control plane certificates for more information.

  • It is recommended that you use Ignition config files within 12 hours after they are generated because the 24-hour certificate rotates from 16 to 22 hours after the cluster is installed. By using the Ignition config files within 12 hours, you can avoid installation failure if the certificate update runs during installation.

Installing the OpenShift CLI by downloading the binary

You can install the OpenShift CLI (oc) to interact with OKD from a command-line interface. You can install oc on Linux, Windows, or macOS.

If you installed an earlier version of oc, you cannot use it to complete all of the commands in OKD 4.12. Download and install the new version of oc.

Installing the OpenShift CLI on Linux

You can install the OpenShift CLI (oc) binary on Linux by using the following procedure.

Procedure

  1. Navigate to https://mirror.openshift.com/pub/openshift-v4/clients/oc/latest/ and choose the folder for your operating system and architecture.

  2. Download oc.tar.gz.

  3. Unpack the archive:

    $ tar xvf <file>

  4. Place the oc binary in a directory that is on your PATH.

    To check your PATH, execute the following command:

    $ echo $PATH
Verification

  • After you install the OpenShift CLI, it is available using the oc command:

    $ oc <command>

Installing the OpenShift CLI on Windows

You can install the OpenShift CLI (oc) binary on Windows by using the following procedure.

Procedure

  1. Navigate to https://mirror.openshift.com/pub/openshift-v4/clients/oc/latest/ and choose the folder for your operating system and architecture.

  2. Download oc.zip.

  3. Unzip the archive with a ZIP program.

  4. Move the oc binary to a directory that is on your PATH.

    To check your PATH, open the command prompt and execute the following command:

    C:\> path
Verification

  • After you install the OpenShift CLI, it is available using the oc command:

    C:\> oc <command>

Installing the OpenShift CLI on macOS

You can install the OpenShift CLI (oc) binary on macOS by using the following procedure.

Procedure

  1. Navigate to https://mirror.openshift.com/pub/openshift-v4/clients/oc/latest/ and choose the folder for your operating system and architecture.

  2. Download oc.tar.gz.

  3. Unpack and unzip the archive.

  4. Move the oc binary to a directory on your PATH.

    To check your PATH, open a terminal and execute the following command:

    $ echo $PATH
Verification

  • After you install the OpenShift CLI, it is available using the oc command:

    $ oc <command>

Logging in to the cluster by using the CLI

You can log in to your cluster as a default system user by exporting the cluster kubeconfig file. The kubeconfig file contains information about the cluster that is used by the CLI to connect a client to the correct cluster and API server. The file is specific to a cluster and is created during OKD installation.

Prerequisites

  • You deployed an OKD cluster.

  • You installed the oc CLI.

Procedure

  1. Export the kubeadmin credentials:

    $ export KUBECONFIG=<installation_directory>/auth/kubeconfig (1)
    1 For <installation_directory>, specify the path to the directory that you stored the installation files in.

  2. Verify you can run oc commands successfully using the exported configuration:

    $ oc whoami
    Example output
    system:admin
Additional resources
Additional resources

Next steps