KumoScale Deployment Modes Available in 3.20

KumoScale version 3.20 supports new deployment modes to meet customer needs and ensure backward compatibility.

KumoScale 3.20 offers three types of deployment:

Appliance Mode: This is the same configuration supported in 3.19 and earlier.  KumoScale software is installed on an appliance and managed as a storage node within the KumoScale Control Cluster.

Managed Mode using an external Kubernetes control cluster:   In this configuration, KumoScale storage node software is installed on your OS and the KumoScale control cluster is your own highly available Kubernetes cluster.

Managed Mode on bare-metal (without Kubernetes): In this configuration, KumoScale storage node software is installed on your OS and the KumoScale Provisioner is deployed as an rpm without any Kubernetes control structure.

How to Determine Which Deployment Mode is Best for Your Environment

Below are the primary features and differences between each deployment mode to  help you determine which mode and configuration to deploy in your environment.


Activity Appliance Mode Managed Mode with Kubernetes Managed Mode on Bare-Metal
Installation The Kubernetes control cluster is installed from a golden image that includes all KumoScale software.

Customers must :
- Install and compile the KumoScale storage node software  for their exact kernel. The storage node software is the same for both types of Managed Mode configurations.

-Install the KumoScale CLI.

-Install and configure the Kubernetes control cluster. This must be a healthy, highly available cluster configured per best practices.

Customers must :
- Install and compile the KumoScale storage node software  for their exact kernel. The storage node software is the same for both types of Managed Mode configurations.

-Install the KumoScale CLI.

-Install the KumoScale Provisioner rpm.

Control Operators and Provisioning KumoScale Operators and the KumoScale Provisioner, provisioner pods, are installed automatically with installation.

- KumoScale Operators must be installed manually.

- The Provisioner pods must be installed and configured manually on the external Kubernetes cluster.

There is no support for control operators. The Provisioner is installed as an rpm and the  Provisioner REST API is used for provisioning storage.

Network Settings

Network settings are configured using KumoScale operators. Users can choose to configure :

-Team

-VLAN

-BGP portal

-Interface settings (MTU, status)

Customers are responsible for configuring network settings before initializing storage nodes.

KumoScale Operators and the REST API support setting the portal only.

Customers are responsible for configuring network settings before initializing storage nodes.

The REST API supports setting the portal only.

Storage Nodes Storage nodes are part of the KumoScale cluster. Nodes can join as master or worker nodes. The KumoScale storage node is NOT part of the Kubernetes cluster. It is a custom resource managed through KumoScale operators and the Provisioner service running on an external Kubernetes cluster. Storage node software is deployed as rpms and all storage node provisioning and management is done using the REST API.
Time Settings KumoScale supports  the time settings (NTP, time zone). It is determined by using the KumoScale operators. Time settings are the customer's responsibility and can not be set or modified through KumoScale operators or the REST API. Time settings are the customer's responsibility and can not be set or modified through the REST API.