Failover to standby in geo-redundant systems for OpenShift and Kubernetes plugins

Note: When the Fabric Services System control plane is in read-only mode or is unreachable, health monitoring fails. The FSC plugin puts itself in the "monitor" state. In this scenario:
  • All Workload Connectivity Intent add or update operations are accepted, but in the failed state. The CRD status is Sync-Add-Failed to reflect the failure. All delete operations are allowed. CRDs can be deleted.
  • All Workload Interface CRD add and update operations are rejected and the CNI must retry. All delete operations are allowed. CRDs can be deleted.
  • When the Fabric Services System control plane is reachable and health monitoring is restored, FSC initiates an audit to reconcile with the Fabric Services System.

Use the following steps to switchover to the standby system for a Kubernetes cluster operating a Fabric Services System managed fabric.

  1. Initiate the switch-over to the new active cluster.

    For instructions, see Initiating failover: switching between the active and standby clusters in the Fabric Services System User Guide.

  2. When the new active cluster is operational, initiate a Connect Full audit on the new active Fabric Services system cluster for the FSC deployment.
    Send the following Connect REST API request:
    HTTP(S) POST {connect_uri}/rest/connect/api/v1/admin/audits {"deploymentId": "<deployment_id>", "scope": "FULL", "dryrun": false}