Get a production cloud with OpenStack Autopilot
Canonical’s OpenStack Autopilot handles every aspect of your production cloud across multiple physical machines through installation, expansion and everyday operations.
- A minimum of 6 machines with the following roles:
- 1 machine for the MAAS server
- 1 machine for the Juju controller
- 1 machine for the OpenStack Autopilot
- 3 or more machines for the cloud (6 or more if you want HA):
- At least one must have 2 NICs
- At least 3 must have 2 disks
- A dedicated switch to create a private cloud LAN
- Internet access through a router on that LAN
Before you start, you’ll want this eBook
The phase change from traditional, monolithic software to multi-host microservices-based big software demands that you approach the challenge of deployment, integration and operations from a different perspective.
This eBook will give you a deeper understanding into why there is a perceived complexity to the installation and operations of OpenStack and how tools like Canonical’s conjure-up and OpenStack Autopilot can help you build a modern, scalable, repeatable and affordable private cloud infrastructure.
1Set up your hardware
Install Ubuntu Server 16.04 LTS on the machine designated to be the MAAS server.
You need to setup a private network with all machines plugged in and enough IP addresses available for all physical and virtual machines you plan to run. This network must not have a DHCP server: MAAS will fill in that role.
For the simplest topology, connect the second NIC of the dual-nic machine(s) to the same network.
To install MAAS, start off on your Ubuntu Server 16.04 LTS machine and type the command below following the step-by-step instructions:
Create your admin credentials and optionally import your SSH key by typing:
Login to the MAAS UI at
Fill in the details on the welcome page and import images for Ubuntu 14.04 LTS and 16.04 LTS. Importing images may take a while, but you can click “continue” as soon as it’s started
If you haven’t imported your SSH key before, now is your chance to import one from Launchpad, GitHub or you can also upload it directly
3Configure the MAAS subnet and DHCP
Go to the “Subnets” tab and click on your subnet. Verify that “gateway” and “DNS” are filled in and update as necessary
Go back (or click on “subnets” again) and this time click on the “untagged” VLAN. Select “Provide dhcp” in the “Take action” button and select a suitable Dynamic range. To start with, count one IP per NIC connected to the network
4Verify image syncing
Go to the “Images” tab and check if the Ubuntu images have all been downloaded and are in a “Synced” state. This is the process started earlier, and depending on your bandwidth it may take a while for it to finish. You can only proceed with the next steps if the images are synced
5Register your hardware with MAAS
Now you need to enlist and commission machines:
- Ensure all machines are set to PXE boot, if possible disable all other boot options, including local disk, in the BIOS
- Power the machines on. They will all appear in the “Nodes” tab of MAAS after a while
- Edit each machine, filling in the power type and other parameters, if they are not correct already
- Select all the machines and, using the “Take action” dropdown, “Commission” them
- Wait until all machines have a “Ready” status
- Verify the networking by going to the details page for the node(s) that have multiple NICs and check that the second NIC (the non-PXE one):
- Is connected to the subnet
- Has the “IP address” field set to “unconfigured”
- The first NIC should be the same except the IP address field will be set to “Auto assign”
6Launch OpenStack Autopilot
Setup Landscape and launch the OpenStack Autopilot with the following commands:
<hostname>above with the hostname of the machine that was selected for the Juju controller role. Do not add the domain name to it
Select “maas” as the cloud type
- Fill in your:
- MAAS server IP
- MAAS user API key — found in your user’s settings in the MAAS user interface (top-right corner)
The next page will show you the list of applications that make up a Landscape deployment. In the case of Landscape, all applications will be deployed to one machine only using LXD containers. If you want to select which machine that will be, you can click on the “Architect” button of any application. Otherwise, click on “Deploy” on the bottom right
When everything is installed, you will be asked to provide details about the Landscape administrator to be created: name, email and password
Finally, a summary will be displayed with a link for you to access Landscape and begin your Openstack deployment. Open that link to access Landscape
Login with your Admin Email and Password
7Review your checklist
At the bottom of the setup page there is a checklist with the status of all of your resources. These should all be green, if it isn’t follow the instructions to resolve.
8Choose your OpenStack components
9Select the hardware on which to deploy the cloud
10Select “Install” to start building your cloud
11Create an OpenStack account to access your Horizon dashboard
12Monitor your region and scale out