Setting Up ONAP¶
ONAP is deployed using the ONAP Operations Manager (OOM).
The recommended ONAP deployment is based on Kubernetes, Docker containers and Helm installer.
The following is the recommended component version.
Software Version Kubernetes 1.11.5 Helm 2.9.1 kubectl 1.11.5 Docker 17.03.x
- The ONAP full installation is validated with the following footprint:
- 14 VM (1 Rancher, 13 K8s nodes) - 8 vCPU - 16 GB RAM
- 160 GB Storage
A quick-start Kubernetes and ONAP Installation Script is available that fully automates the instructions in the installation guides below to spin up an HA-enabled Kubernetes cluster and deploys ONAP on it using OOM.
Creation of Kubernetes cluster is described here:
ONAP installation is described here:
NodePorts are used to allow client applications, that run outside of Kubernetes, access to ONAP components deployed by OOM. A NodePort maps an externally reachable port to an internal port of an ONAP microservice. It should be noted that the use of NodePorts is temporary. An alternative solution is currently being scoped for the Dublin Release.
- The list of node ports used by ONAP is documented in ONAP Node Port List.
Prior to deployment of ONAP, there is no need to download manually any Docker container. The OOM deployment takes care to automatically download the Docker containers.
It is also possible to deploy a subset of ONAP components on a single VM. The VM flavor to be used depends on the number of ONAP components to be deployed.
For test purposes, it is possible to use other installations: