Installing the Assistant chat by using a Helm chart

You must deploy Assistant chat agents to a Red Hat OpenShift cluster by using a Helm chart in the IBM Cloud Container Registry package. You use a values.yaml file to populate the required fields for your environment.

1. Log in to the IBM Cloud Container Registry from any machine that is connected to the OCP cluster and replace <ZDBA_ENTITLEMENT_KEY> with your IBM Cloud API key:

   echo -n "<ZDBA_ENTITLEMENT_KEY>" | helm registry login icr.io \
     --username iamapikey \
     --password-stdin

2. Pull and extract the chart package. The latest version is 1.1.0.1, for example: --version 1.1.0-1.

   helm pull oci://icr.io/z-database-assistant/z-database-assistant-charts --version 1.1.0-1 --untar

3. Change to the directory that contains the values.yaml file, the deploy.sh file, and the templates/ subdirectory before you continue.

   cd z-database-assistant-charts

4. Open the values.yaml file and populate the fields that are empty or need environment-specific values before you deploy.
   Work with your OCP administrator or the person responsible for maintaining your watsonx Assistant™ for Z to obtain these values.

   Restriction: The values.yaml file can contain sensitive values. Do not commit environment-specific secrets to version control.

   General settings

   The following settings configure the language model runtime for your deployment.

   entitlementKey

   IBM Cloud entitlement key used for the image pull secret. This key is the same key that you use to pull the Helm chart.

   env.modelRuntime

   Model runtime type. You must set this value to cloud, openai-protocol, or on-prem. The value you specify determines which model runtime settings are required, as described in the following three sections. Complete the section that applies to the model runtime time that you specified.

   Settings required for the cloud model runtime type

   The following settings are required if you set the env.modelRuntime setting to cloud.

   env.projectId or env.deploymentSpaceId

   Set either the env.projectId or the env.deploymentSpaceId field depending on whether you are working with a project or deployment space in watsonx.ai. These parameters are mutually exclusive.

   env.llmBaseUrl

   Base URL for the model service endpoint.

   env.llmApiKey

   API key for the configured language model service.

   Settings required for the openai-protocol model runtime type

   The following settings are required if you set the env.modelRuntime setting to openai-protocol.

   env.llmModel

   Model name or identifier to use for inference, for example, openai/gpt-oss-120b.

   env.llmBaseUrl

   Base URL for the model service endpoint.

   env.llmApiKey

   API key for the configured language model service.

   Settings required for the on-prem model runtime type

   The following settings are required if you set the env.modelRuntime setting to on-prem.

   env.llmModel

   Model name or identifier to use for inference, for example, openai/gpt-oss-120b.

   env.llmBaseUrl

   Base URL for the model service endpoint.

   env.cpdUsername

   Cloud Pak for Data user name if applicable, for example, cpdadmin.

   env.cpdVersion

   Cloud Pak for Data version if applicable, for example, 5.3.

   env.instanceId

   Instance identifier. The default is openshift.

   env.llmApiKey

   API key for the configured language model service.

   Values from watsonx Assistant for Z

   The following settings connect IBM® Z Database Assistant to watsonx Assistant for Z services for knowledge retrieval and document ingestion.

   env.opensearchUser

   User name for the OpenSearch wrapper service.

   env.opensearchPassword

   Password for the OpenSearch wrapper service.

   env.opensearchRoute

   Route or URL for the OpenSearch wrapper service.

   env.ingestionServiceUrl

   URL for the ingestion service used for knowledge file ingestion.

   env.ingestionServicePassword

   Password for the ingestion service.

5. Review and set these fields appropriately if your environment requires client certificate authentication for the IMS Operations API:

   env.mcpCertSecret.cer

   Client certificate for IMS Operations API authentication. This field contains placeholder content that must be replaced if certificate authentication is required.

   env.mcpCertSecret.key

   Private key for the client certificate.

   env.mcpCertSecret.serverCA

   Server CA certificate that is used to validate the remote endpoint.

6. Deploy to OpenShift.
   1. Because the deploy.sh script requires a .env file for OpenShift cluster credentials, create the .env file from the .env.example file by running the following command:

      cp .env.example .env
      

   2. Enable the file permissions for the .env file and the deploy.sh script by running the following commands:

      
      chmod 600 .env
      chmod +x deploy.sh

   3. Edit the .env file and populate it with your environment-specific values. These are the required values from the .env.example file:

      OCP_CLUSTER

      OCP cluster API endpoint, not the OCP cluster web endpoint.

      OCP_USERNAME

      User name that has access to the OCP cluster to deploy the agents.

      OCP_PASSWORD

      Password for that user name.

      Save and close the .env file.

   4. Run the deploy.sh script to deploy the Assistant chat agents to the OpenShift cluster:

      
      ./deploy.sh -i

      Tip: For more information about the deploy.sh script file, run ./deploy.sh -h.

./deploy.sh -r >> install-log.txt