OOM Cloud Setup Guide¶
OOM deploys and manages ONAP on a pre-established Kubernetes cluster - the creation of this cluster is outside of the scope of the OOM project as there are many options including public clouds with pre-established environments. However, this guide includes instructions for how to create and use some of the more popular environments which could be used to host ONAP. If creation of a Kubernetes cluster is required, the life-cycle of this cluster is independent of the life-cycle of the ONAP components themselves. Much like an OpenStack environment, the Kubernetes environment may be used for an extended period of time, possibly spanning multiple ONAP releases.
Inclusion of a cloud technology or provider in this guide does not imply an endorsement.
The versions of Kubernetes that are supported by OOM are as follows:
Minimum Hardware Configuration¶
The hardware requirements are provided below. Note that this is for a full ONAP deployment (all components). Customizing ONAP to deploy only components that are needed will drastically reduce the requirements.
|224GB||160GB||112||0.0.0.0/0 (all open)|
Kubernetes supports a maximum of 110 pods per node - configurable in the –max-pods=n setting off the “additional kubelet flags” box in the kubernetes template window described in ‘ONAP Development - 110 pod limit Wiki’ - this limit does not need to be modified . The use of many small nodes is preferred over a few larger nodes (for example 14x16GB - 8 vCores each). Subsets of ONAP may still be deployed on a single node.
OOM can be deployed on a private set of physical hosts or VMs (or even a combination of the two). The following guide describe the recommended method to setup a Kubernetes cluster: ONAP on HA Kubernetes Cluster.
There are alternative deployment methods described on the Cloud Native Deployment Wiki