Designing an ARM-based Enterprise Cloud

Ampere is a semiconductor design company that develops server-grade processors for cloud environments, with a focus on high performance and energy efficiency.

You can deploy and verify an OpenNebula cloud on Ampere hardware, by using Certified Ampere Hardware with OpenNebula software, a set of Ansible playbooks that allows you to deploy and verify an OpenNebula cloud with a few simple commands.

This guide provides a reference Ampere hardware specification that has been used to verify OpenNebula. It includes instructions on how to perform a ZeroTouch deployment of OpenNebula on the certified hardware, and provides a reference architecture and configuration.

Following this guide, you can:

• Perform a Zero-touch deployment of OpenNebula over these resources.
• Ensure the correct operation of the resulting cloud using an automated verification procedure.

Additionally, this guide includes a brief description of how to instantiate a Virtual Machine, to help you get started on your OpenNebula Cloud.

Basic Outline of the Deployment Procedure

Performing the deployment involves these high-level steps:

1. Clone the dedicated OpenNebula on Ampere GitHub repository on your deployment machine.
2. Modify the repository with the parameters for your Ampere servers.
3. Perform the automated deployment to the Ampere servers.
4. Verify the deployment by running the automated verification command.

Additional Information Resources

• Certified Ampere Hardware with OpenNebula software
• Ampere Resource Library

Hardware Specification and Architecture

This section contains the specification of the used Ampere hardware and software resources for the reference OpenNebula deployment.

Architecture

The target high-level cloud architecture overview is shown below. Two Ampere servers are used: the first for hosting the OpenNebula Front-end services and VMs, the second for hosting VMs only. A simple VXLAN networking is configured for the communication between the VMs, so all deployed VMs are attached to the same logical LAN.

Hardware Specification

The tables below detail the characteristics for the Front-end, virtualization host, storage, networking and provisioning model.

Front-end Requirements

Two identical servers were used. One of them acts as Front-end and both of them acts as hosts.

Host Requirements

Storage Specification

Network Requirements

Provisioning Model

Server Specifications

Automated Deployment and Configuration

To perform the automated deployment of an OpenNebula cloud, the Ampere infrastructure of connected servers must be previously configured and available. This guide provides guidance on how to extract the list of required parameters of the provisioned infrastructure – which will later be used for the automation of the OpenNebula deployment – and an outline of the process for the automated deployment of an OpenNebula cloud.

Ampere Infrastructure Provisioning

Provisioning the Ampere infrastructure on premises is out of the scope of this guide. To perform the automated deployment of an OpenNebula cloud, the Ampere infrastructure must meet the following conditions:

• Servers are provisioned
• Networking is configured
• Storage is configured
• The required operating system is installed
• Servers in the infrastructure are reachable from the machine where the deployment commands will be run

For the reference architecture and HW/SW specifications, please refer to the Hardware Specification and Architecture section.

Save Required Parameters

To proceed with OpenNebula deployment, we need to extract and save some required parameters that the deployment automation relies on.

Deployment and Automated Verification Procedure

The complete OpenNebula deployment procedure and all of the required resources are available in the Certified Hardware Ampere for OpenNebula, also referred to as the deployment repository. For instructions on how to use the required parameters extracted from the provisioned Ampere servers, please check the README file in the repo.

The deployment procedure consists of the following high-level steps:

1. Clone the deployment repository.
2. Update the deployment repository with the required parameters gathered above.
3. Launch the deployment automation commands.
4. Launch the verification automation command.

Validate Certified Hardware Deployments

After successfully verifying the infrastructure deployed by the automations, to run a Virtual Machine access the OpenNebula web UI at:

http://<Front-end IP>:2616/fireedge/sunstone

To log in, use the default username oneadmin, and the password specified in the one_pass variable of the inventory file.

The image below shows the Alpine Linux 3.20 Virtual Machine included in the OpenNebula installation:

To instantiate the VM, in the Sunstone UI’s left-hand menu go to Instances –> VMs. Click the Create icon highlighted below, then select the Virtual Machine template. Follow the steps of the VM instantiation wizard. For this basic guide, all values can be left empty or at their defaults.

In the Instances -> VMs view select the new VM in the list to open the details page Configuration tab. Select Update configuration.

In the Context tab of the modal dialog that opens, scroll down to the Context Custom Variables section and expand it. Find the PASSWORD field and select Update (the pencil icon), then specify the desired root password for the VM, then click Accept, as shown below, then press Continue.

Log in to the VM via VNC, go to the ellipsis drop-down menu and select Console -> VNC. Log in as user root with the password that you specified in the previous step.

After accessing the deployed VM’s command line interface, verify that the terminal is responsive. For example, change to the home folder of user root:

root@vm:~# cd ~
root@vm:~# pwd
/root

Finally, as a cleanup step, terminate the VM by clicking the red “Trash can” icon, then verify that the VM transitions to state DONE, as shown below.