Installing IBM watsonx.ai

An instance administrator can install IBM watsonx.ai.

Who needs to complete this task?

Instance administrator To install IBM watsonx.ai, you must be an instance administrator. An instance administrator has permission to install software in the following projects:

The operators project for the instance

The operators for this instance of IBM watsonx.ai are installed in the operators project.

In the installation commands, the ${PROJECT_CPD_INST_OPERATORS} environment variable refers to the operators project.

The operands project for the instance

The custom resources for the control plane and IBM watsonx.ai are installed in the operands project.

In the installation commands, the ${PROJECT_CPD_INST_OPERANDS} environment variable refers to the operands project.

When do you need to complete this task?

Review the following options to determine whether you need to complete this task:

If you want to install multiple services at the same time, follow the process in Running a batch installation of solutions and services instead.
If you didn't install IBM watsonx.ai as part of a batch installation, complete this task to add IBM watsonx.ai to your environment.
Repeat as needed If you are responsible for multiple instances of IBM® Software Hub, you can repeat this task to install more instances of IBM watsonx.ai on the cluster.

Information you need to complete this task

Review the following information before you install IBM watsonx.ai:

Version requirements: All of the components that are associated with an instance of IBM Software Hub must be installed at the same release. For example, if the IBM Software Hub control plane is installed at Version 5.3.1, you must install IBM watsonx.ai at Version 5.3.1.

Environment variables

The commands in this task use environment variables so that you can run the commands exactly as written.

If you don't have the script that defines the environment variables, see Setting up installation environment variables.
To use the environment variables from the script, you must source the environment variables before you run the commands in this task. For example, run:
```
source ./cpd_vars.sh
```

Security context constraint: IBM watsonx.ai works with the default Red Hat® OpenShift® Container Platform security context constraint, restricted-v2.

Common core services

IBM watsonx.ai requires the IBM Software Hub common core services.

If the common core services are not installed in the operands project for the instance, the common core services are automatically installed when you install IBM watsonx.ai. The common core services installation increases the amount of time the installation takes to complete.

Storage requirements

You must specify storage classes when you install IBM watsonx.ai. The following storage classes are recommended. However, if you don't use these storage classes on your cluster, ensure that you specify a storage class with an equivalent definition.

Storage	Notes	Storage classes
OpenShift Data Foundation	When you install the service, specify file storage and block storage.	File storage: `ocs-storagecluster-cephfs` Block storage: `ocs-storagecluster-ceph-rbd`
IBM Fusion Data Foundation	When you install the service, specify file storage and block storage.	File storage: `ocs-storagecluster-cephfs` Block storage: `ocs-storagecluster-ceph-rbd`
IBM Fusion Global Data Platform	When you install the service, specify the same storage class for both file storage and block storage.	File storage: Either of the following storage classes: `ibm-spectrum-scale-sc` `ibm-storage-fusion-cp-sc` Block storage: Either of the following storage classes: `ibm-spectrum-scale-sc` `ibm-storage-fusion-cp-sc`
IBM Storage Scale Container Native	When you install the service, specify the same storage class for both file storage and block storage.	File storage: `ibm-spectrum-scale-sc` Block storage: `ibm-spectrum-scale-sc`
Portworx	When you install the service, the `--storage_vendor=portworx` option ensures that the service uses the correct storage classes.	File storage: `portworx-rwx-gp3-sc` (Equivalent to `portworx-shared-gp3` in older installations) Block storage: `portworx-couchdb-sc` `portworx-elastic-sc` `portworx-gp3-sc`
NFS	When you install the service, specify the same storage class for both file storage and block storage.	File storage: `managed-nfs-storage` Block storage: `managed-nfs-storage`
Amazon Elastic storage	When you install the service, specify file storage. If you specify block storage, the service ignores this information. File storage is provided by Amazon Elastic File System.	File storage: `efs-nfs-client`
Nutanix	Not supported.	Not applicable.

Before you begin

This task assumes that the following prerequisites are met:

System requirements

This task assumes that the cluster meets the minimum requirements for IBM watsonx.ai.

Where to find more information
If this task is not complete, see System requirements.

In addition, ensure that you have the appropriate type and number of GPU for IBM watsonx.ai.

Where to find more information
If this task is not complete, see GPU requirements.

Workstation

This task assumes that the workstation from which you will run the installation is set up as a client workstation and has the following command-line interfaces:

IBM Software Hub CLI: cpd-cli
OpenShift CLI: oc
Helm CLI: helm

Where to find more information
If this task is not complete, see Setting up a client workstation.

Control plane

This task assumes that the IBM Software Hub control plane is installed.

Where to find more information
If this task is not complete, see Installing an instance of IBM Software Hub.

Private container registry

If your environment uses a private container registry (for example, your cluster is air-gapped), this task assumes that the following tasks are complete:

The IBM watsonx.ai software images are mirrored to the private container registry.
Where to find more information

If this task is not complete, see Mirroring images to a private container registry.
The cpd-cli is configured to pull the olm-utils-v4 image from the private container registry.
Where to find more information

If this task is not complete, see Pulling the olm-utils-v4 image from the private container registry.

Load balancer

Where to find more information
If this task is not complete, see Mirroring images to a private container registry.

Where to find more information
If this task is not complete, see Pulling the olm-utils-v4 image from the private container registry.

The load balancer timeout settings are adjusted for IBM watsonx.ai.

Where to find more information
If this task is not complete, see Changing load balancer settings.

GPU operators

This task assumes that the operators required to use GPUs are installed.

Where to find more information
If this task is not complete, see Installing operators for services that require GPUs.

Red Hat OpenShift AI

This task assumes that Red Hat OpenShift AI is installed.

Where to find more information
If this task is not complete, see Installing Red Hat OpenShift AI.

Cluster-scoped resources

This task assumes that the cluster-scoped resources, such as custom resource definitions, cluster roles, and cluster role bindings, exist.

Where to find more information
If this task is not complete, see Creating cluster-scoped resources for the IBM Software Hub platform and services.

Image pull secrets

This task assumes that the secrets that contain the image pull credentials for the instance exist.

Where to find more information
If this task is not complete, see Creating image pull secrets for an instance of IBM Software Hub.

Prerequisites for hosting foundation models

Private container registry requirements: When you complete the steps in Mirroring images directly to a private container registry, you can specify group names to mirror only images for the foundation models that you plan to use. The group name for each foundation model is provided in the tables that list the supported foundation models. See System requirements for foundation models in IBM watsonx.ai.
GPU requirements: To inference large language models, extra compute resources are needed, including specialized GPUs. Make sure that your cluster meets the system requirements for the models you want to use and that the required operators to support GPUs are installed before you install the IBM watsonx.ai service. Again, refer to System requirements for foundation models in IBM watsonx.ai for more information about specific foundation model resource needs. After you complete the steps for Installing operators for services that require GPUs, you can optionally configure NVIDIA Multi-Instance GPU (MIG). However, do not partition GPU processors in a cluster that you plan to use for tuning foundation models. For more information, see Partitioning GPU processors in IBM watsonx.ai.

Procedure

Complete the following tasks to install IBM watsonx.ai:

Specifying installation components
Specifying installation options
Installing the service
Validating the installation
What to do next

Specifying installation components

You can install multiple components with the watsonx.ai™ service. Set the XAI_COMPONENT_TYPE environment variable to the list of components you want to install.

Choose from the following components to install with the service:

watsonx_ai component

Installs the core watsonx.ai functionality that lets you work with foundation models installed in your cluster and use tools such as the Prompt Lab and Tuning Studio.

Set the XAI_COMPONENT_TYPE as follows to install only the core watsonx.ai component:

export XAI_COMPONENT_TYPE=watsonx_ai

model_gateway component

Installs the model gateway that lets you work with foundation models that are hosted on a remote cluster or third-party infrastructure.

Set the XAI_COMPONENT_TYPE as follows to install only the model gateway component:

export XAI_COMPONENT_TYPE=model_gateway

Note: The model gateway does not require GPUs or Red Hat OpenShift AI.

To install both the watsonx.ai core functionality and the model gateway, set the XAI_COMPONENT_TYPE as follows:

export XAI_COMPONENT_TYPE=watsonx_ai,model_gateway

Specifying installation options

If you plan to install watsonx.ai, you can specify the installation options in a file named install-options.yml in the cpd-cli work directory (For example: cpd-cli-workspace/olm-utils-workspace/work).

Important: The installation options apply to the watsonx_ai core component only. These options do not apply to the model_gateway component.

The parameters are optional. If you do not set these installation parameters, the default values are used.

Follow the appropriate guidance for the version of IBM Software Hub that you installed:

Version 5.3.1

The sample YAML content uses the default values.

5.3.1 and later The formatting applies only to IBM Software Hub Version 5.3.1.

Retain the --- syntax at the beginning of the entry to ensure that this entry is treated as a separate document.

---
# ............................................................................
# watsonx.ai parameters
# ............................................................................
non_olm:
  watsonxAi:
    liteInstall: false

Version 5.3.0

The sample YAML content uses the default values.

If you want to override one or more of the default values, add the parameters to the non_olm: section of the install-options.yml file under the watsonxAi: entry.

# ............................................................................
# watsonx.ai parameters
# ............................................................................
  watsonxAi:
    liteInstall: false

Property	Description
`liteInstall`	Specify whether you want to install the full watsonx.ai service or the watsonx.ai lightweight engine. For more information, see Choosing an IBM watsonx.ai installation mode. Default value `false` Valid values `false` Install the full service. `true` Install the lightweight engine.

Installing the service

To install IBM watsonx.ai:

Log the cpd-cli in to the Red Hat OpenShift Container Platform cluster:
```
${CPDM_OC_LOGIN}
```
Remember: CPDM_OC_LOGIN is an alias for the cpd-cli manage login-to-ocp command.

Install the operator and custom resource for IBM watsonx.ai.

The command that you run depends on the storage on your cluster.

Red Hat OpenShift Data Foundation storage

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

IBM Fusion Data Foundation storage

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

IBM Fusion Global Data Platform storage

Remember: When you use IBM Fusion Global Data Platform storage, both ${STG_CLASS_BLOCK} and ${STG_CLASS_FILE} point to the same storage class, typically ibm-spectrum-scale-sc or ibm-storage-fusion-cp-sc.

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

IBM Storage Scale Container Native storage

Remember: When you use IBM Storage Scale Container Native storage, both ${STG_CLASS_BLOCK} and ${STG_CLASS_FILE} point to the same storage class, typically ibm-spectrum-scale-sc.

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

Portworx storage

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--storage_vendor=portworx \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--storage_vendor=portworx \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

NFS storage

Remember: When you use NFS storage, both ${STG_CLASS_BLOCK} and ${STG_CLASS_FILE} point to the same storage class, typically managed-nfs-storage.

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

AWS with EFS storage only

Remember: When you use EFS storage, both ${STG_CLASS_BLOCK} and ${STG_CLASS_FILE} point to the same storage class, typically efs-nfs-client.

Run the appropriate command for your environment:

Default installation (without installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET}

Custom installation (with installation options)

cpd-cli manage install-components \
--license_acceptance=true \
--components=${XAI_COMPONENT_TYPE} \
--release=${VERSION} \
--patch_id=${PATCH_ID} \
--operator_ns=${PROJECT_CPD_INST_OPERATORS} \
--instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--block_storage_class=${STG_CLASS_BLOCK} \
--file_storage_class=${STG_CLASS_FILE} \
--image_pull_prefix=${IMAGE_PULL_PREFIX} \
--image_pull_secret=${IMAGE_PULL_SECRET} \
--param-file=/tmp/work/install-options.yml

Validating the installation

IBM watsonx.ai is installed when the install-components command returns:

[SUCCESS]... The install-components command ran successfully

If you want to confirm that the custom resource status is Completed, you can run the cpd-cli manage get-cr-status command:

cpd-cli manage get-cr-status \
--cpd_instance_ns=${PROJECT_CPD_INST_OPERANDS} \
--components=${XAI_COMPONENT_TYPE}

The following services are automatically installed when you install the IBM watsonx.ai service:

Watson Studio
Watson Machine Learning

What to do next

Install foundation models that you want to work with in IBM watsonx.ai. For details, see Post-installation setup for IBM watsonx.ai. IBM watsonx.ai is now ready to use.

Tip: The watsonx.ai lightweight engine does not have a dedicated user interface. You can use the IBM Software Hub web client to perform administrative tasks such as monitoring platform resource usage. For details about how to access the web client, see Installing the required components for an instance of IBM Software Hub.