Good morning everyone, this is Alejandro. And today we are going to talk about a project that i did with my colleges when i was working at IBM.
The project was about to create a Living Lab in a state university. When i said Living Lab, i am talking about a space where we can create projects that will help our city to become a Smarter City.
So, my team, was selected to configure an IBM cluster with the technology of OpenStack. But at the beginning, the team and i were like: What the hell is Openstack? So for the beginners, OpenStack is a set of tools for managing a cloud computing platform, this tool can be used for public or private use. And i have to say that OpenStack is built by the community and is open source. Also, another companies like Red Hat and RackSpace give support to the platform.
When you install OpenStack in a computer or a cluster, you will have to define the “nodes” that you are going to use. We have typically three kinds of “nodes”, the first is the Controller Node, the second is the Computer Node and finally the Storage Node.
So now, we know what OpenStack is, but we have a lot of concepts that we barely know or never hear about them. So i would like to explain some of them.
- Controller Node. The controller node is where you are going to have the brain of the stack. In this node you will control everything and in almost every component that you install you will have to install keys in the controller node.
- Compute Node. The compute node is where you are going to have all the power to execute the tasks. In this node is where you are going to take the hardware, for my experience in this part i highly recommend to put more RAM and if it is possible more CPU.
- Storage Node. The storage node is divided in two. The name of the components are Cinder and Swift (we are going to talk about later). Here you can have the two storage nodes. The first one it can be Cinder. Cinder is built to store info and be able to access that information. And the second is Swift. Swift is built for backups of the information in your storage node. It is optimized to look for data among files and folders.
Right now, we have the definition and some basic concepts that you have to get familiar. And i did this because, i think that if i have found the information like this it would be more easy to me, to understand the process of how to install and how to manage the cluster.
Another concepts that from my perspective are essential are:
- Horizon. Is the service that we are going to install, to have a dashboard were we can create, modify or delete virtual machines, users, networks, etc. Also we can gather information about the hardware that is in use.
- Neutron. Is the service that is going to offer the network capability for OpenStack. So the services could communicate between them.
- Nova. It is the primary service and is the one that will help us to create virtual machines.
- Keystone. Is like a security layer, because is the one that grants the permissions to the services and the users.
- Glance. Is the one that provides images services to OpenStack. The images are virtual copies of the hard disks. These images can be used as templates when you deploy a new virtual machine.
So, my experience with this kind of tool at the beginning was awful, believe me. The team was frustrated, because we did not find enough information. And for the installation, we have a lot of errors and mistakes. So my advice is if you want to learn about Openstack, first learn about Linux, basic concepts in the command line. Next, know your hardware, network and your goals. And finally, be patient, go to the OpenStack and use a lot Google, believe me. Google is going to be your best friend.
The period of the project at the beginning was estimated for 2 months. 1 day per week. But this was not enough so we started to go 2 o 3 days per week. Also, this was not our only project, we had like 3 projects more. Then, is something that i was forgotten and is the training that you are going to give to your users. I mean, the users needs to have basic experience with Virtual Machines.
As i said we had a lot of problems, some of them because we did not know what to do and others that was from the university. But at the end, we completed the project. Only for the installation of Openstack took us like 6 months.