The compute service in Nova is considered to be the core component of OpenStack. Understanding how to scale out the workload among several identical compute nodes might require to briefly revisit the building blocks of Nova:

• nova-compute: This runs on the compute node as described in Chapter 1, Inflating the OpenStack Setup. It is responsible for communicating with the hypervisors. Nova-compute interacts with each hypervisor by means of drivers. It creates compute resources by picking up requests from the message queue.
• nova-scheduler: This runs on the cloud controller as described in Chapter 1, Inflating the OpenStack Setup. It is responsible for finding the right placement (physical server) of the initiated request to create a VM. The request will be left in the message queue along with additional information regarding the server information where...

In this section, we will add a new compute host to our first initial OpenStack environment using Ansible. The new compute node will have similar hardware requirements to the first compute node, as follows:

• Memory: At least 4 GB RAM
• Processor: At least 4 64-bit x86 CPUs
• Disk space: At least 40 GB free disk space
• Network: At least two NICs

The following excerpt will be added to the vagrant file in our test environment:

    ...
    # Compute Node
    config.vm.define :cn02 do |cn02|
      cn02.vm.hostname= "cn02"
      cn02.vm.provider "virtualbox" do |vb|
              vb.customize ["modifyvm", :id, "--memory", "4096"]
              vb.customize ["modifyvm", :id, "--cpus", "4"]
                 vb.customize ["modifyvm", :id, "--nicpromic2", &quot...

One of the key successes of OpenStack is its exposure to other hypervisor vendors—no vendor lock-in. As virtualization platforms have continued to evolve over the last decade, cloud providers have faced a major challenge regarding how to unlock their opportunities to keep growing. OpenStack makes it possible to support and integrate more virtualization technologies at pace—more freedom of choice. Extending your private OpenStack cloud should not only be limited to bringing more power to the compute fleet, but also enabling flexibility across supported virtualization platforms. At the time of writing, OpenStack supports the following hypervisor technologies:

• Kernel-based Virtual Machine (KVM) (libvirt)
• QEMU
• LXC
• XenServer
• VMware
• Docker
• Hyper-V

What about letting your OpenStack cloud environment go beyond its limits? This can be the right moment to revisit the design of your private cloud when it grows in size. Most importantly, an operational team should be aware that the offered service should keep functioning continuously. To become more efficient as the user's resource base grows, one should pinpoint more specifically the worker blocks of the cloud—compute nodes. Again, OpenStack comes with great concepts that would help to massively scale your compute power. This is where the art of segregation comes. Segregation in OpenStack can be classified into two main umbrellas:

• Infrastructure segregation: This involves logical grouping based on conceptual and physical capabilities. Typically, it defines a specific number of clusters organized to scale horizontally.
• Workload segregation...

The OpenStack compute service leverages many concepts and terminologies described in this chapter in order to enable additional OpenStack capabilities. We have learned how adding a new compute node can be performed from code using Ansible. More advanced layouts have been exposed to take advantage of several supported hypervisors by OpenStack. We have covered a few of them, such as Docker and Xen. VMware and HyperV are also powerful candidates to support Nova hypervisors in OpenStack, but they were not covered. This chapter demonstrates not only that adopting a hybrid hypervisor setup might expand your OpenStack environment, but also that you might expand it adopting compute segregation. It is a big plus to design a compute cluster setup by taking into consideration how they should be grouped locally and spread widely. We have covered a logical grouping layout using host...