Upgrading the Platform Software

Prerequisites

• You have performed all the tasks described in Before Upgrading the Platform. You copy the upgrade bundle to the correct location as part of the pre-check process.

• Required access rights: Platform Administrator

  The Platform Administrator must have login access to the Controller host as either root or the user who performed the original installation, as appropriate.

• You have upgraded HPE Ezmeral Data Fabric on Kubernetes, if applicable.

• All hosts are powered on and accessible.

Upgrade Process Overview

• Each of the hosts in will be upgraded. If the upgrade fails on one or more hosts, then the entire upgrade process will be rolled back on all hosts.

• The upgrade process updates the Controller host near the beginning of the upgrade process. At this point, the management service will be momentarily interrupted in order to change over to the new version; thus, explicitly refreshing the web UI during this interval may result in a browser error. If this occurs, refresh the page after a brief wait in order to get the latest version of the page.

• If you have applied a custom-authentication-bundle to userconfig.tgz file as described in Modifying the Authentication Package, back up the userconfig.tgz file before proceeding with platform upgrade process.

For information about the impact of the upgrade operation on existing workloads, see Upgrading to HPE Ezmeral Runtime Enterprise 5.7.x.

After you complete this task, you perform other upgrade tasks, depending on what you have implemented in your environment.

Upgrade Procedure

1. Enter Lockdown mode via the web interface by opening the Quick Access menu and then selecting .

   The Enter system lockdown dialog appears.

2. Enter a descriptive reason for the lockdown in the Enter Reason field, and then enter Lockdown mode by clicking Submit.

3. In the Available Upgrades tab of the Installation screen, click the Upgrade button for the build that you want to upgrade to.

   Lockdown mode will be confirmed, upgrade compatibility with your current version will be confirmed, and installation will then either proceed or display an error message.

   The Confirm Upgrade popup appears.

4. Enter a brief note in the Notes field if you like. Any notes you enter will be saved and can be viewed later.
5. (Optional) Select the Auto-Roll check box. Enabling this option upgrades the Controller hosts first, one at a time. After all the controller hosts are upgraded successfully, the upgrade is committed, and then the remaining hosts in the deployment are upgraded in batches. After the upgrade is committed, you can continue using HPE Ezmeral Runtime Enterprise while the remaining hosts are being upgraded.
6. Click Proceed to continue the upgrade.

7. Monitor the status of the upgrade process:

   The Upgrade Progress section appears on the Available Upgrades tab to display the status of the upgrade progress. Status values are the following:

   • Pending: The upgrade has not yet started.
   • Upgrading: The upgrade is taking place. Additional details appear during this phase, as HPE Ezmeral Runtime Enterprise extracts the upgrade package, upgrades the Controller host and Worker hosts, and finishes the upgrade. The package version being applied also appears.
   • Finalizing: Post-upgrade cleanup is occurring.
   • Complete: The upgrade has completed successfully.
   • Rolling Back: Upgrade has encountered an error and is reverting HPE Ezmeral Runtime Enterprise back to the original version.
   • Error: The upgrade did not complete successfully.

8. Monitor the status of the upgrade of the nodes. You can check the status of the nodes at any time during the upgrade process.

   To show the status of the nodes, click the Show Upgrade Details and Actions link to display the Upgrade Details and Actions window.

   Per-node status values should be interpreted in the same way as the status values of the overall upgrade process. There is also an additional Upgraded status, which indicates that the individual node has upgraded successfully while the overall upgrade process continues on other nodes.

   • After a host has a status of Upgraded, click the Commit Upgrade button to commit the HPE Ezmeral Runtime Enterprise upgrade on the hosts that successfully upgraded.

   • If the upgrade fails before the Primary Controller host is updated, then the message Upgrade Failed will appear and you will see the status error displayed for the primary host. You will need to investigate the Primary Controller host to resolve any issues, and then try again.

   • If the upgrade fails on another host after the Primary Controller has been upgraded, then the message Awaiting manual recovery will appear and you will see the status error displayed for the failed hosts. You will need to investigate the affected hosts to resolve any issues, and then try again.
9. If errors occur during the upgrade, do the following:

   • Check the affected host and resolve issues, then retry installation on the host by clicking Retry Host.

     The message Retrying appears in the Available Upgrades tab.

   • After a host has a status of Upgraded, commit the HPE Ezmeral Runtime Enterprise upgrade on the host by clicking Commit Upgrade.

   To cancel the upgrade process on all hosts and revert to the previously-installed version of HPE Ezmeral Runtime Enterprise, Click Cancel and Rollback Upgrade.

   You can view the completed upgrade in the Upgrade History tab of the Installation screen. .

10. (Optional) If you had applied a custom-authentication-bundle to userconfig.tgz file before the platform upgrade as described in Modifying the Authentication Package, do the following steps:
    • Restore the backed up the userconfig.tgz file to /opt/bluedata/catalog/postconfig/ after the upgrade.
    • In the HPE Ezmeral Runtime Enterprise web UI, access the Platform Authentication page as Site Admin, and click Submit (without changing any settings). This will update /etc/sssd/sssd.conf file in all EPIC Apps, with the SSSD configuration generated by the customized authentication bundle.
11. Do one of the following:
    • If your environment uses Kubernetes:
      NOTE

      Before you can create or edit Kubernetes clusters, You must exit Lockdown Mode.

      • If you choose to install the Falco Kernel Modules, after the platform upgrade is completed, you can proceed to (Optional) Installing Falco Kernel Modules on Hosts
      • If you choose not to install the Falco Kernel Modules, proceed to Verifying the Upgrade.
    • If your environment uses EPIC for container management, proceed to Verifying the Upgrade.