Preparing for the upgrade

Unified Compute Platform (UCP) Advisor Software Installation Guide

Version
4.5.x
Audience
anonymous
Part Number
MK-92UCP120-13
Perform the following procedure on each of the following VMs:
  • Master Node VM
  • Worker Node1 VM
  • Worker Node2 VM
  1. Take a snapshot of each VM and verify that the datastore has approximately 3x200 GB of additional available space to manage the snapshots. The total available disk space in the datastore for all three VMs should be approximately 2 TB.
  2. Verify that there is SSH connectivity to each VM to ensure that all the VMs are up and running.
  3. Verify that each VM meets its respective resource requirements. See UCP Advisor resource requirements.
  4. Verify that only the Master Node VM and the SCP server are configured with their respective IP addresses. No other custom IP addresses should be configured.
  5. Verify that the Worker Node1 and Worker Node2 VMs have a single network IP address.
  6. Verify that the required ports on each VM are not blocked by the firewall. See Firewall configuration required by the management node.
  7. From the Hitachi Support Connect Portal (login required), download the UCP Advisor 4.5 upgrade package ucpadvisor-air-gapped-v4.5.0-88.tar.gz to the /home/ucpadmin folder on the UCP Advisor Master Node VM.
  8. Log on to the Master Node VM as a ucpadmin user.
  9. Navigate to the /home/ucpadmin folder, and delete any previous installer folder, if present.
  10. Extract the upgrade package. Enter:
    tar -xvzf ucpadvisor-air-gapped-v4.5.0-88.tar.gz
  11. Relogin to the Master Node VM as a root user, and then navigate to the /home/ucpadmin/installer/scripts/ folder.
  12. Set permissions for the ucpadmin user to perform the upgrade. Enter the following commands:

    For Oracle Linux, enter:

    chmod 777 set_permissions_preupgrade.sh
    ./set_permissions_preupgrade.sh --worker=<Worker Node1 IP address> --worker=<Worker Node2 IP address>

    For CentOS, enter:

    chmod 777 set_permissions_preupgrade_centos.sh
    ./set_permissions_preupgrade_centos.sh --worker=<Worker Node1 IP address> --worker=<Worker Node2 IP address>
  13. Enter the root password when prompted.
  14. Navigate to the /root folder on the Master Node and Worker Node VMs, and delete the os_upgrade folder, if present.
  15. Relogin to the Master Node VM as a ucpadmin user.
  16. Download the upgrade package for your operating system from the Hitachi Support Connect Portal (login required), and then copy it to the UCP Advisor Master Node VM:
    • CentOS 7.9: ucpa_os_upgrade_pkg_centos_4.5-1117.tar.gz
    • Oracle Linux 8.6: ucpa_os_upgrade_pkg_4.5-1117.tar.gz
  17. Navigate to the /home/ucpadmin folder on the Master Node VM and run the following commands to update each VM with the CentOS 7.9 or Oracle Linux 8.6 update:

    For CentOS 7.9:

    tar –xvzf ucpa_os_upgrade_pkg_centos_4.5-1117.tar.gz
    cd os_upgrade
    sudo ./upgrade_rpm --master-node <Master Node IP address> --worker-nodes <Worker Node1 IP address> <Worker Node2 IP address> --node-username <Node username> --node-password <Node password>
    Note: You can safely ignore any NOKEY warnings during the CentOS 7.9 package update. They do not affect the update.

    For Oracle Linux 8.6:

    tar –xvzf ucpa_os_upgrade_pkg_4.5-1117.tar.gz
    cd os_upgrade
    ./upgrade_rpm --master-node <Master Node IP address> --worker-nodes <Worker Node1 IP address> <Worker Node2 IP address> --node-username <Node username> --node-password <Node password>
    After the VMs are successfully updated, the following message is displayed:
    RPMs installed on master node <Master Node VM IP address>
    RPMs installed on worker node <Worker Node1 VM IP address>
    RPMs installed on worker node <Worker Node2 VM IP address>
    Time taken xxx Minutes
  18. If you are upgrading UCP Advisor to version 4.5 from a base 4.0 or 4.1 environment, you must update the Kubernetes package. See Updating Kubernetes package.
  19. (Optional) Generate and download the log bundle and any reports that you want to keep.
  20. (Optional) Identify the node for the command device. Enter the following command on the Master Node VM:
    kubectl get node -o wide -l app=ucp
    Note: This command displays a value that points to a Worker Node VM.
  21. (Optional) Provision a new command device or update the existing command device on the Worker Node VM obtained in step 17. See Provisioning a command device for the UCP Advisor Worker Node VM.
  22. If you have an external Gateway VM configured, then you must update it to the latest version before upgrading the UCP Advisor management software. See Updating the external Gateway VM.