OpenStack – Setup on Digital Ocean
This article discusses our quick OpenStack setup in support of our first CompatibleOne cloud broker setup. Read on to see what we did to run a simple VM within OpenStack.
Overview
CompatibleOne has some online documentation that discusses running a first VM within OpenStack. So, we want to setup OpenStack using our Digital Ocean account.
We selected the OpenStack 3-Node Architecture consisting of a Controller, Network, and Compute node.
- All Nodes. CentOS 6.5 x64.
- Controller Node. 2GB Ram, 40GB SSD Disk, Private Networking.
- Network Node. 512MB Ram, 20GB SSD Disk, Private Networking.
- Compute Node. 2GB Ram, 40GB SSD Disk, Private Networking.
Basic Setup
Prior to doing anything else, perform these steps:
- Create local SSH group and a local user. We created group
ssh-usersand userl.abruce. We madel.abrucea member ofusers, wheel, ssh-usersgroups. - Setup
/etc/sudoers. We used the easiest option of%wheel ALL=(ALL) NOPASSWD: ALL. - Disable
rootlogin. We modified/etc/ssh/sshd_configwith the following:# 9-MAY-14, ABr AllowGroups ssh-users PermitRootLogin noBe sure to verify that you can login after you use
service sshd restart🙂
Setup Networking
OpenStack configuration guide assumes eth1 is Management, eth2 is Instance Tunnels, and eth3 is external. Digital Ocean only gives you two NICs: eth1 is your external interface and eth2 is your private interface. So we will lose the IP address for eth2 and use sub-interfaces for each network. No need for a VLAN!
Team-oriented systems mentor with deep knowledge of numerous software methodologies, technologies, languages, and operating systems. Excited about turning emerging technology into working production-ready systems. Focused on moving software teams to a higher level of world-class application development. Specialties:Software analysis and development...Product management through the entire lifecycle...Discrete product integration specialist!
This is a very timely tutorial; but where is the rest of the article?
Yes, Al – I have been remiss on this one! It came down to cost; DO is cheap compared to EC2 and Azure (1/3 cost) *but* you still get charged whether you have machines on or off. When I realized that, I already had a bigger balance than I planned for. So the machines had to go off 🙂
I have some beater servers that I will eventually get provisioned and see if I can run KVM on them. That will keep the cost down. Anyways, in the meantime I have a bunch more work to do on setting up puppet and getting it automated. Plus some 60-day trials to burn on Rackspace / etc. So maybe I can get the OS work fired back up again 🙂