API nodes for IBM Storage Scale native REST API

The IBM Storage Scale native REST API feature includes a new node class called API nodes. The native REST API is available only through the API nodes, which expose the full suite of APIs provided by this feature. API nodes also support local commands through the scalectl CLI without the --url option. By default, API nodes are quorum nodes.

API nodes manage and replicate metadata that is required for the native REST API, including long-running operations (LRO) data. These nodes can ensure high availability of the API. Because quorum is required for IBM Storage Scale to function, quorum nodes are designated as API nodes by default. The following figure illustrates the architecture of a cluster with three quorum or API nodes:
Figure 1. High-level architecture of a cluster with API nodes
High-level architecture of a cluster with API nodes

The native REST API server does not start on non-API nodes. These nodes expose only a subset of native REST API functions through scalectl commands, which are limited to managing local configurations, such as node identities.

Only the following scalectl commands are supported locally on non-API nodes:
  • scalectl node config
  • scalectl cluster
  • scalectl config

To use additional native REST API functions from a non-API nodes, issue the scalectl command with the --url <FQDN/IP>:<port> flag. The FQDN or IP must point to an API node, and the request is sent over HTTPS.

Note:
  • If the native REST API feature is enabled and the quorum designation of a node is changed, restart scaleadmd service on nodes where the quorum designation was changed.