We recently received a user question to clarify the process of adding instances to an HCI cluster via the Command-Line Interface (CLI) and how to properly structure the deployment configuration in the request.
To add a new instance, install HCI on the new instance(s), and run hci_setup with the IP addresses of the existing master instances of your cluster. When you start HCI on these new instances, they will automatically join the cluster, but by default, new instances will have no services running on them.
The general idea is (for a fresh install):
For instance additions/removals:
For service reconfiguration:
The ServiceUpdateModel is used across API operations and can be used to specify the configuration of one or more services to perform update actions, where each service is represented by a ServiceConfigModel. Not all of the fields in ServiceConfigModel are required for all API calls, and deltas for specific services are supported. Only the serviceId, serviceName, and action are required in these objects. See the REST API documentation in the UI for details.
Retrieving data ...