Controllers Tab

The Controllers tab of the Controllers/Upgrades screen (see The Controllers/Upgrades Screen) allows the Platform Administrator to enable or disable platform High Availability protection, which protects against the failure of a single host (Primary Controller, Shadow Controller, or Arbiter).

Note: Enabling platform High Availability protection does not protect any virtual clusters that are running. Please see High Availability for information about cluster vs. platform protection and Creating a New Cluster, Creating a New Training Cluster, Creating a New Notebook Cluster, or Creating a New Deployment Cluster for more information on enabling cluster-level protection.

Please see the following:

Enabling HA Protection
Disabling HA Protection

Enabling HA Protection

Note: HPE recommends enabling platform High Availability before adding a large number of Kubernetes or EPIC hosts.

To enable platform High Availability:

Ensure that all of the requirements listed in High Availability Requirements are met.
If you have any existing virtual clusters, then those clusters must be deleted and then recreated after enabling High Availability.
Enter site lockdown as described in Lockdown Mode.
Check the Enable HA check box.
Proceed as follows:
- If the Controller and Standby Controller hosts are in different subnets, then you must leave the Cluster IP field blank. By leaving both the Cluster IP and Cluster Name fields blank, you may access the web interface by navigating to http://<gateway_ip> or https://<gateway_ip>, as appropriate, where <gateway_ip> is the IP address of a Gateway host. See Gateway Hosts.
- If the Controller and Standby Controller hosts are on the same subnet, then you can enter an available IP address to use as the cluster IP address in the Cluster IP field. This IP address must be in the same subnet as the Controller host and cannot be in use by any other resource. If you do not supply a cluster IP address, then you may access the web interface by navigating to http://<gateway_ip> or https://<gateawy_ip>, as appropriate, if you have defined a cluster name in the Cluster Name field. This cluster name must be mapped to the cluster IP address via a user-accessible DNS server.
Note: For same-subnet scenario, the external switch connecting the hosts to the network must support gratuitous arp in order for the cluster IP address to function correctly.

Note: After enabling HA, you may use either the cluster IP address or cluster name to log into the web interface, because this will automatically connect you to the Controller host (during normal operation) or the Shadow Controller host (when HA protection has been triggered by Controller host failure). If the Controller host fails, then you will not be able to access the web interface using the IP address of that host.
Select the hosts to use as the Shadow Controller, and Arbiter Node using the Shadow Controller and Arbiter Node pull-down menus. If the deployment has three hosts, then the host remaining after assigning the Shadow Controller host will be the Arbiter, and vice versa. If there are more than three hosts, then you may select one of the remaining Worker hosts to be the Arbiter after selecting the Shadow Controller, and vice-versa. You cannot remove or modify the Shadow Controller or Arbiter host after enabling High Availability protection.
Note: Hosts that did not have internal storage defined when being added as described in Step 4: Select Hard Drives cannot be used as a Shadow Controller or Arbiter.
Click Submit. The Controllers tab displays the message HA Setup in progress. This process may take up to 30 minutes to complete depending on a number of factors. If desired, you may click the Details button to open the HA Setup Details popup, which provides detailed information about the HA setup process.
A message appears in the upper right corner of the web interface once the process completes informing you have HPE Ezmeral Container Platform is now running in High Availability mode and reminding you to begin using the cluster IP address or name that you entered in Step 4 to log in to the web interface going forward. Clicking the Click here to migrate to Cluster Name link in this message logs you out of the web interface and returns you to the Login screen using the cluster IP address.
Log into the web interface using your normal username and password.
Exit site lockdown as described in Lockdown Mode.
Create any needed virtual cluster(s).

If enabling High Availability fails, then the fields in the Controllers tab will reappear, and the deployment will continue running in its previous non-High Availability state. Please contact HPE Technical Support for assistance.

Note: If you installed the Network Manager service while installing the base operating system on the hosts, then this service will stop because it conflicts with the High Availability monitoring services.

Note: If you are using ISSI, then be to use a mechanism (e.g. Puppet, Chef, etc.) that will replicate the ISSI configuration to the Shadow Controller host. See Implementation-Specific Script Interface.

Disabling HA Protection

If platform HA is enabled, then you may disable it as follows:

Enter site lockdown as described in Lockdown Mode.
Clear the Enable HA button. A confirmation popup appears. Click OK to confirm.
Click Submit to begin disabling HA protection. You will be automatically logged out and then migrated to the Controller IP address.
Log in to the web interface via the Controller IP address using your normal username and password. The HA tab will display the message HA disable in progress. Disabling HA may require up to 30 minutes to complete. If desired, you may click the Details button to open the HPE HA Disable Details popup, which provides additional information about the HA disable process.
After the HA disable process finishes, exit site lockdown as described in Lockdown Mode.

Disabling platform HA protection has the following effects:

The Shadow Controller and Arbiter hosts become Worker hosts.
The cluster IP address is disabled; you must log in using the Controller IP address
The HPE Container Platform platform is no longer protected against Controller host failure.

Note: You may change the hosts used for the Shadow Controller and Arbiter roles by disabling HA protection and then re-enabling it and using the updated IP addresses/hostnames.

Note: Please contact HPE Technical Support if disabling HA fails.

Platform HA can be disabled under one of the following situations:

All hosts are online: See All Hosts Online.
One or more host(s) are offline: See Some Hosts Offline.

All Hosts Online

If all hosts are online, then HA protection will start being disabled when the Platform Administrator submits this task, and the deployment will be placed into a non-HA state.

Some Hosts Offline

If one or more host(s) are offline, then he disable HA task will be rejected, and the Platform Administrator will be notified that one or more host(s) are offline. This can happen for one of several reasons:

Worker offline: If a Worker host is offline, then that host will retain its HA configuration when it comes back up, which will not be consistent with the platform status. You can simply delete the Worker host as described in Decommissioning/Deleting an EPIC Worker and then repeat the HA disable process.
Shadow Controller failure: Normally, disabling HA protection will delete the database replicas on the Shadow Controller and Arbiter hosts. If the Shadow Controller host is offline due to a hardware failure, then database deletion will fail. When the faulty hardware is replaced and HA protection is re-enabled, the original Arbiter host should be re-designated as the Arbiter, and the new Shadow Controller host must use the same IP address as the previous Shadow Controller host.
Arbiter failure: Normally, disabling HA protection will delete the database replicas on the Shadow Controller and Arbiter hosts. If the Arbiter host is offline due to a hardware failure, then database deletion will fail. When the faulty hardware is replaced and HA protection is re-enabled, the original Shadow Controller host should be re-designated as the Shadow Controller, and the new Arbiter host must use the same IP address as the previous Arbiter host.

Please contact HPE Technical Support for assistance if you need to disable HA protection while the Shadow Controller or Arbiter host is down. The Support team will perform some manual operations to allow the management service to ignore the offline host and allow you to proceed with disabling HA Protection. When you replace the hardware, be sure to follow the instructions listed in the previous bullets, as applicable.