Chapter 2. Preparing to deploy Red Hat Decision Manager in your OpenShift environment

Before deploying Red Hat Decision Manager in your OpenShift environment, you must complete several tasks. You do not need to repeat these tasks if you want to deploy additional images, for example, for new versions of decision services or for other decision services

2.1. Ensuring the availability of image streams and the image registry

To deploy Red Hat Decision Manager components on Red Hat OpenShift Container Platform, you must ensure that OpenShift can download the correct images from the Red Hat registry. To download the images, OpenShift requires image streams, which contain the information about the location of images. OpenShift also must be configured to authenticate with the Red Hat registry using your service account user name and password.

Some versions of the OpenShift environment include the required image streams. You must check if they are available. If image streams are available in OpenShift by default, you can use them if the OpenShift infrastructure is configured for registry authentication server. The administrator must complete the registry authentication configuration when installing the OpenShift environment.

Otherwise, you can configure registry authentication in your own project and install the image streams in that project.

Procedure

  1. Determine whether Red Hat OpenShift Container Platform is configured with the user name and password for Red Hat registry access. For details about the required configuration, see Configuring a Registry Location. If you are using an OpenShift Online subscription, it is configured for Red Hat registry access.
  2. If Red Hat OpenShift Container Platform is configured with the user name and password for Red Hat registry access, enter the following commands:

    $ oc get imagestreamtag -n openshift | grep -F rhdm76-decisioncentral-openshift
    $ oc get imagestreamtag -n openshift | grep -F rhdm76-kieserver-openshift

    If the outputs of both commands are not empty, the required image streams are available in the openshift namespace and no further action is required.

  3. If the output of one or both of the commands is empty or if OpenShift is not configured with the user name and password for Red Hat registry access, complete the following steps:

    1. Ensure you are logged in to OpenShift with the oc command and that your project is active.
    2. Complete the steps documented in Registry Service Accounts for Shared Environments. You must log in to the Red Hat Customer Portal to access the document and to complete the steps to create a registry service account.
    3. Select the OpenShift Secret tab and click the link under Download secret to download the YAML secret file.
    4. View the downloaded file and note the name that is listed in the name: entry.
    5. Enter the following commands:

      oc create -f <file_name>.yaml
      oc secrets link default <secret_name> --for=pull
      oc secrets link builder <secret_name> --for=pull

      Replace <file_name> with the name of the downloaded file and <secret_name> with the name that is listed in the name: entry of the file.

    6. Download the rhdm-7.6.0-openshift-templates.zip product deliverable file from the Software Downloads page and extract the rhdm76-image-streams.yaml file.
    7. Enter the following command:

      $ oc apply -f rhdm76-image-streams.yaml
      Note

      If you complete these steps, you install the image streams into the namespace of your project. In this case, when you deploy the templates, you must set the IMAGE_STREAM_NAMESPACE parameter to the name of this project.

2.2. Creating the secrets for Decision Server

OpenShift uses objects called secrets to hold sensitive information such as passwords or keystores. For more information about OpenShift secrets, see the Secrets chapter in the OpenShift documentation.

You must create an SSL certificate for HTTP access to Decision Server and provide it to your OpenShift environment as a secret.

Procedure

  1. Generate an SSL keystore with a private and public key for SSL encryption for Decision Server. For more information on how to create a keystore with self-signed or purchased SSL certificates, see Generate a SSL Encryption Key and Certificate.

    Note

    In a production environment, generate a valid signed certificate that matches the expected URL for Decision Server.

  2. Save the keystore in a file named keystore.jks.
  3. Record the name of the certificate. The default value for this name in Red Hat Decision Manager configuration is jboss.
  4. Record the password of the keystore file. The default value for this name in Red Hat Decision Manager configuration is mykeystorepass.
  5. Use the oc command to generate a secret named kieserver-app-secret from the new keystore file:

    $ oc create secret generic kieserver-app-secret --from-file=keystore.jks

2.3. Creating the secrets for Business Central

You must create an SSL certificate for HTTP access to Business Central and provide it to your OpenShift environment as a secret.

Do not use the same certificate and keystore for Business Central and Decision Server.

Procedure

  1. Generate an SSL keystore with a private and public key for SSL encryption for Business Central. For more information on how to create a keystore with self-signed or purchased SSL certificates, see Generate a SSL Encryption Key and Certificate.

    Note

    In a production environment, generate a valid signed certificate that matches the expected URL for Business Central.

  2. Save the keystore in a file named keystore.jks.
  3. Record the name of the certificate. The default value for this name in Red Hat Decision Manager configuration is jboss.
  4. Record the password of the keystore file. The default value for this name in Red Hat Decision Manager configuration is mykeystorepass.
  5. Use the oc command to generate a secret named decisioncentral-app-secret from the new keystore file:

    $ oc create secret generic decisioncentral-app-secret --from-file=keystore.jks

2.4. Preparing a Maven mirror repository for offline use

If your Red Hat OpenShift Container Platform environment does not have outgoing access to the public Internet, you must prepare a Maven repository with a mirror of all the necessary artifacts and make this repository available to your environment.

Note

You do not need to complete this procedure if your Red Hat OpenShift Container Platform environment is connected to the Internet.

Prerequisites

  • A computer that has outgoing access to the public Internet is available.

Procedure

  1. Prepare a Maven release repository to which you can write. The repository must allow read access without authentication. Your OpenShift environment must have access to this repository. You can deploy a Nexus repository manager in the OpenShift environment. For instructions about setting up Nexus on OpenShift, see Setting up Nexus. Use this repository as a separate mirror repository.

    Alternatively, if you use a custom external repository (for example, Nexus) for your services, you can use the same repository as a mirror repository.

  2. On the computer that has an outgoing connection to the public Internet, complete the following steps:

    1. Download the latest version of the Offliner tool.
    2. Download the rhdm-7.6.0-offliner.txt product deliverable file from the Software Downloads page of the Red Hat Customer Portal.
    3. Enter the following command to use the Offliner tool to download the required artifacts:

      java -jar offliner-<version>.jar -r https://maven.repository.redhat.com/ga/ -r https://repo1.maven.org/maven2/ -d /home/user/temp rhdm-7.6.0-offliner.txt

      Replace /home/user/temp with an empty temporary directory and <version> with the version of the Offliner tool that you downloaded. The download can take a significant amount of time.

    4. Upload all artifacts from the temporary directory to the Maven mirror repository that you prepared. You can use the Maven Repository Provisioner utility to upload the artifacts.
  3. If you developed services outside Business Central and they have additional dependencies, add the dependencies to the mirror repository. If you developed the services as Maven projects, you can use the following steps to prepare these dependencies automatically. Complete the steps on the computer that has an outgoing connection to the public Internet.

    1. Create a backup of the local Maven cache directory (~/.m2/repository) and then clear the directory.
    2. Build the source of your projects using the mvn clean install command.
    3. For every project, enter the following command to ensure that Maven downloads all runtime dependencies for all the artifacts generated by the project:

      mvn -e -DskipTests dependency:go-offline -f /path/to/project/pom.xml --batch-mode -Djava.net.preferIPv4Stack=true

      Replace /path/to/project/pom.xml with the correct path to the pom.xml file of the project.

    4. Upload all artifacts from the local Maven cache directory (~/.m2/repository) to the Maven mirror repository that you prepared. You can use the Maven Repository Provisioner utility to upload the artifacts.

2.5. Changing GlusterFS configuration

You must check whether your OpenShift environment uses GlusterFS to provide permanent storage volumes. If it uses GlusterFS, to ensure optimal performance of Business Central, you must tune your GlusterFS storage by changing the storage class configuration.

Procedure

  1. To check whether your environment uses GlusterFS, enter the following command:

    oc get storageclass

    In the results, check whether the (default) marker is on the storage class that lists glusterfs. For example, in the following output the default storage class is gluster-container, which does list glusterfs:

    NAME              PROVISIONER                       AGE
    gluster-block     gluster.org/glusterblock          8d
    gluster-container (default) kubernetes.io/glusterfs 8d

    If the result has a default storage class that does not list glusterfs or if the result is empty, you do not need to make any changes. In this case, skip the rest of this procedure.

  2. To save the configuration of the default storage class into a YAML file, enter the following command:

    oc get storageclass <class-name> -o yaml >storage_config.yaml

    Replace <class-name> with the name of the default storage class. Example:

    oc get storageclass gluster-container -o yaml >storage_config.yaml
  3. Edit the storage_config.yaml file:

    1. Remove the lines with the following keys:

      • creationTimestamp
      • resourceVersion
      • selfLink
      • uid
    2. If you are planning to use Business Central only as a single pod, without high-availability configuration, on the line with the volumeoptions key, add the following options:

      features.cache-invalidation on
      performance.nl-cache on

      For example:

      volumeoptions: client.ssl off, server.ssl off, features.cache-invalidation on, performance.nl-cache on

    3. If you are planning to use Business Central in a high-availability configuration, on the line with the volumeoptions key, add the following options:

      features.cache-invalidation on
      nfs.trusted-write on
      nfs.trusted-sync on
      performance.nl-cache on
      performance.stat-prefetch off
      performance.read-ahead off
      performance.write-behind off
      performance.readdir-ahead off
      performance.io-cache off
      performance.quick-read off
      performance.open-behind off
      locks.mandatory-locking off
      performance.strict-o-direct on

      For example:

      volumeoptions: client.ssl off, server.ssl off, features.cache-invalidation on, nfs.trusted-write on, nfs.trusted-sync on, performance.nl-cache on, performance.stat-prefetch off, performance.read-ahead off, performance.write-behind off, performance.readdir-ahead off, performance.io-cache off, performance.quick-read off, performance.open-behind off, locks.mandatory-locking off, performance.strict-o-direct on

  4. To remove the existing default storage class, enter the following command:

    oc delete storageclass <class-name>

    Replace <class-name> with the name of the default storage class. Example:

    oc delete storageclass gluster-container
  5. To re-create the storage class using the new configuration, enter the following command:

    oc create -f storage_config.yaml

2.6. Provisioning persistent volumes with ReadWriteMany access mode using NFS

If you want to deploy high-availability Business Central, your environment must provision persistent volumes with ReadWriteMany access mode.

Note

If you want to deploy a high-availability authoring environment, for optimal performance and reliability, provision persistent volumes using GlusterFS. Configure the GlusterFS storage class as described in Section 2.5, “Changing GlusterFS configuration”.

If your configuration requires provisioning persistent volumes with ReadWriteMany access mode but your environment does not support such provisioning, use NFS to provision the volumes. Otherwise, skip this procedure.

Procedure

Deploy an NFS server and provision the persistent volumes using NFS. For information about provisioning persistent volumes using NFS, see the "Persistent storage using NFS" section of the Configuring Clusters guide.