Table of contents

Install Data Science Experience Local

Requirement: Before you install DSX Local, ensure that your servers meet all of the system requirements.

You can install by web browser UI or by configuration file. Restrictions: The Amazon Web Services (AWS) installation supports a command line or configuration file installation only. A load balancer can only be specified in a configuration file installation.

Install by web browser UI

To install DSX Local by web browser UI, complete the following steps:

  1. Download the DSX Local installation package to the first master node of your dedicated DSX Local cluster. Ensure that the master node has a public IP address that is reachable by your web browser server.

  2. Move the installation package to the installer files partition you created (you cannot install from any path that is mounted as /), and make the installation package executable.

  3. Optional: To automatically install add-ons such as SPSS Modeler or H2O Flow, create a new addon-packages sub-directory in the same directory as the installer files. Download the installation package TAR files into the addon-packages sub-directory. Later, the installer will prompt for confirmation to install those add-ons.

  4. In /etc/selinux/config, ensure SELINUX=enforcing or SELINUX=permissive for DSX Local to install.

  5. Run the installation package with the --extract-pre-install-scripts parameter to extract the pre-installation scripts for DSX Local. RHEL cluster only: You can automatically install an RHEL docker on each node (requires a minimum of 200 GB raw disk) by completing the following steps on each node in the cluster:

    1. Enable a Red Hat repo for package installation on each node.
    2. Enable the extras repo on each node so docker can be installed, for example, --subscription-manager repos --enable rhel-7-server-extras-rpms.
    3. Ensure that there is a raw disk available (200G or more) on each node that has no partitions and no file systems. You can use wipefs --all --force /dev/xxx to prepare it.
    4. Run the docker_redhat_install.sh script you extracted on all nodes in the cluster to install docker using the RHEL repo.
  6. If you chose not to use the docker_redhat_install.sh script and selected SELINUX=enforcing, then after installing docker complete the following steps:

    1. In the /etc/sysconfig/docker file, remove --selinux-enabled from the OPTIONS line.
    2. Enter the following command: setsebool -P docker_transition_unconfined 1
    3. Restart docker.
  7. Run the extracted pre-installation check script on each node in the cluster:

    • For a cluster with five nodes or less, run pre_install_check.sh --type=3nodes on all of the nodes.
    • For a cluster that exceeds five nodes, run pre_install_check.sh --type=9nodes_master on all master nodes, pre_install_check.sh --type=9nodes_compute on all compute nodes, and pre_install_check.sh --type=9nodes_deploy on all deployment nodes.
    • The script prompts for information about your installation such as path to the installation files, location of data (for master nodes), and the subnet being used by weave (the default is 9.242.0.0/16). You will get a pass or a fail for mandatory requirements, or a warning for issues such as cron jobs asking to make sure there are no scripts being run that might change the network configuration.
  8. For a six or more node installation by web browser UI, run the installation package without any parameters. For a three, four, or five node installation by web browser UI, run the installation package with the --three-nodes parameter. If you opted to install CentOS packages (including docker) to fulfill the DSX Local prerequisites, include the --centos-repo parameter. Wait several minutes for the pre-installation steps to complete.

  9. Enter W to continue the installation by web browser. Yyou will receive a web address to continue the installation in a web browser).

  10. Agree to the terms and conditions.

  11. In the Assign & ping nodes page, the Private key filename field is optional. You can specify a private SSH key file (instead of a sudo user name and password) to install by using the private key as the credentials (ensure that this key can be used to make a successful SSH connection to each node in your cluster). Alternatively, you can leave the field blank and specify sudo user names and passwords instead.

    Figure: Panel for assigning a minimum of six nodes

    Assign and ping nine nodes screencap

    Figure: Panel for assigning a minimum of three nodes

    Assign and ping three nodes screencap

  12. Specify a cluster overlay network, or type in the suggested value of 9.242.0.0/16. Ensure that the value differs from any networks that exist within your local network. Otherwise, DSX Local encounters conflicts when it tries to access outside resources.

  13. Specify one unused IP address as the proxy IP address. The IP address must be static, portable, and in the same subnet as the cluster.

  14. Specify the IP addresses of the nodes to install the components on. The program determines whether they meet the system requirements. If you are not using a private key file, then specify sudo user names and passwords for each node. Specify the path to store installer files and the path for service data storage (these two paths need to be the partitions pre-created and mounted with the xfs format).

  15. Optional: Add extra nodes to scale your cluster. In a runtime environment, DSX Local supports adding compute and deployment nodes only.

  16. Start the installation. The program installs the Kubernetes cluster across the nodes you specified, configures the network and storage, sets up the docker registries, deploys the services such as Cloudant and Spark, and deploys DSX and its user interfaces. This installation runs step-by-step, and might take a long time. If the page displays improperly due to a network issue, refresh the page. If an installation step fails, you can view the log and contact IBM support. To avoid unexpected issues, make sure to resolve the issue in the log before you retry or skip the step. Tip: If you see an error during deployment of DSX or dashboard helm charts, then do not retry or skip; instead, watch the logs by running the following command: kubectl get po --all-namespaces | grep ibmdp-assistant-deployment | awk '{system("kubectl logs -n "$1" "$2)}'.

    Figure: Log file for a failed installation step with the retry or skip option

    Installation failure screencap

    Figure: Panel for a successful installation

    Installation success screencap

  17. Set the password for your primary DSX administrator (user name is admin).

  18. Click the Submit and Launch button to open the DSX Local client automatically. As the primary DSX administrator, you can toggle to both the DSX Local client and administration dashboard client from the same web address.

    Context switcher for DSX Local client to the admin dashboard

See set up to configure your DSX Local system.

Install by command line and configuration file

Alternatively, you can automate the DSX installation by creating a configuration file that will complete all of the parameters: wdp.conf. The file must be in the same directory as the installation package before you run the installation, and the installation package must be in the installer files partition.

To install DSX Local by configuration file, complete the following steps:

  1. Download the DSX Local installation package to the first master node of your dedicated DSX Local cluster.

  2. Move the installation package to the installer files partition you created (you cannot install from any path that is mounted as /), and make the installation package executable.

  3. Optional: To automatically install add-ons such as SPSS Modeler or H2O Flow, create a new addon-packages sub-directory in the same directory as the installer files. Download the installation package TAR files into the addon-packages sub-directory. Later, the installer will prompt for confirmation to install those add-ons.

  4. In /etc/selinux/config, ensure SELINUX=enforcing or SELINUX=permissive for DSX Local to install.

  5. Run the installation package with the --extract-pre-install-scripts parameter to extract the pre-installation scripts for DSX Local. RHEL cluster only: You can automatically install an RHEL docker on each node (requires a minimum of 200 GB raw disk) by completing the following steps on each node in the cluster:

    1. Enable a Red Hat repo for package installation on each node.
    2. Enable the extras repo on each node so docker can be installed, for example, --subscription-manager repos --enable rhel-7-server-extras-rpms.
    3. Ensure that there is a raw disk available (200G or more) on each node that has no partitions and no file systems. You can use wipefs --all --force /dev/xxx to prepare it.
    4. Run the docker_redhat_install.sh script you extracted on all nodes in the cluster to install docker using the RHEL repo.
  6. If you chose not to use the docker_redhat_install.sh script and selected SELINUX=enforcing, then after installing docker complete the following steps:

    1. In the /etc/sysconfig/docker file, remove --selinux-enabled from the OPTIONS line.
    2. Enter the following command: setsebool -P docker_transition_unconfined 1
    3. Restart docker.
  7. Run the extracted pre-installation check script on each node in the cluster:

    • For a cluster with five nodes or less, run pre_install_check.sh --type=3nodes on all of the nodes.
    • For a cluster that exceeds five nodes, run pre_install_check.sh --type=9nodes_master on all master nodes, pre_install_check.sh --type=9nodes_compute on all compute nodes, and pre_install_check.sh --type=9nodes_deploy on all deployment nodes.
    • The script prompts for information about your installation such as path to the installation files, location of data (for master nodes), and the subnet being used by weave (the default is 9.242.0.0/16). You will get a pass or a fail for mandatory requirements, or a warning for issues such as cron jobs asking to make sure there are no scripts being run that might change the network configuration.
  8. If you already have a wdp.conf file, put it in the same directory as the installation package. To generate a new wdp.conf template, run the installation with parameter --get-conf-user (for a sudo user) or --get-conf-key (for an SSH key).

    • For a six or more node installation, no further parameters are required.
    • For a three, four, or five node wdp.conf template, add the --three-nodes parameter.
    • If you opted to install CentOS packages (including docker) to fulfill the DSX Local prerequisites, add the --centos-repo parameter.
    • For an Amazon Web Services (AWS) installation, add the --aws parameter. Tip: Use the --help parameter for assistance.
  9. After wdp.conf is generated, you can customize the parameters. Ensure that every field is completed. To use your own load balancer for the HA cluster instead of a proxy IP, comment out the virtual_ip_address= line in the wdp.conf file and insert the new line: load_balancer_ip_address=. Requirement: The load balancer must be configured before installing DSX Local to forward traffic for port 6443 and 443 to all three control nodes (or master nodes) with persistent IP round robin for the cluster to function properly.

  10. Run the installation package, and ensure it detects the wdp.conf file. You will be prompted to use this file for the configuration.

  11. Accept the terms and conditions, and type in the password (if you're using the user name). If an installation step fails, you can view the log and contact IBM support. To avoid unexpected issues, make sure to resolve the issue in the log before click before you retry or skip the step. Tip: If you see an error during deployment of DSX or dashboard helm charts, then do not retry or skip; instead, watch the logs by running the following command: kubectl get po --all-namespaces | grep ibmdp-assistant-deployment | awk '{system("kubectl logs -n "$1" "$2)}'.

  12. When the installation completes successfully, save the URL address for the DSX Local client. Sign in as the primary DSX administrator (default user is admin) and change the default password of password. You can toggle to both the DSX Local client and administration dashboard client from the same web address.

    Context switcher for DSX Local client to the admin dashboard

See set up to configure your DSX Local system.

Install for PowerBroker

DSX Local Version 1.2.0.2 or later supports a number of PowerBroker configurations. PowerBroker manages Privileged Access Management in computing units, and is used to log user actions and privilege management through a central server.

To run the installer with pbrun, enter the following command:

pbrun bash -c "export TERM=\"xterm-256color\"; cd /ibm; ./installer_name --pbrun"

where /ibm represents the installation directory.

If you entered a non-root user, the add node pop up displays radio buttons that ask for Become Method. Choose pbrun and continue as normal.

To uninstall using pbrun, enter the following command:

pbrun bash -c "/wdp/utils/uninstall.sh --username=your_user_here --password --pbrun"