kolla-ansible/doc/source/reference/compute/nova-guide.rst

Nova - Compute Service

Nova is a core service in OpenStack, and provides compute services. Typically this is via Virtual Machines (VMs), but may also be via bare metal servers if Nova is coupled with Ironic.

Nova is enabled by default, but may be disabled by setting enable_nova to no in globals.yml.

Virtualisation Drivers

The virtualisation driver may be selected via nova_compute_virt_type in globals.yml. Supported options are qemu, kvm, and vmware. The default is kvm.

Libvirt

Information on the libvirt-based drivers kvm and qemu can be found in libvirt-guide.

VMware

Information on the VMware-based driver vmware can be found in vmware-guide.

Bare Metal

Information on using Nova with Ironic to deploy compute instances to bare metal can be found in ../bare-metal/ironic-guide.

Fake Driver

The fake driver can be used for testing Nova's scaling properties without requiring access to a large amount of hardware resources. It is covered in nova-fake-driver.

Consoles

The console driver may be selected via nova_console in globals.yml. Valid options are none, novnc and spice. Additionally, serial console support can be enabled by setting enable_nova_serialconsole_proxy to yes.

Cells

Information on using Nova Cells V2 to scale out can be found in nova-cells-guide.

Vendordata

Nova supports passing deployer provided data to instances using a concept known as Vendordata. If a Vendordata file is located in the following path within the Kolla configuration, Kolla will automatically use it when the Nova service is deployed or reconfigured: /etc/kolla/config/nova/vendordata.json.

Failure handling

Compute service registration

During deployment, Kolla Ansible waits for Nova compute services to register themselves. By default, if a compute service does not register itself before the timeout, that host will be marked as failed in the Ansible run. This behaviour is useful at scale, where failures are more frequent.

Alternatively, to fail all hosts in a cell when any compute service fails to register, set nova_compute_registration_fatal to true.

Managing resource providers via config files

In the Victoria cycle Nova merged support for managing resource providers via configuration files <admin/managing-resource-providers>.

Kolla Ansible limits the use of this feature to a single config file per Nova Compute service, which is defined via Ansible inventory group/host vars. The reason for doing this is to encourage users to configure each compute service individually, so that when further resources are added, existing compute services do not need to be restarted.

For example, a user wanting to configure a compute resource with GPUs for a specific host may add the following file to host_vars:

[host_vars]$ cat gpu_compute_0001
nova_cell_compute_provider_config:
  meta:
    schema_version: '1.0'
  providers:
    - identification:
        name: $COMPUTE_NODE
      inventories:
        additional:
          - CUSTOM_GPU:
              total: 8
              reserved: 0
              min_unit: 1
              max_unit: 1
              step_size: 1
              allocation_ratio: 1.0

A similar approach can be used with group vars to cover more than one machine.

Since a badly formatted file will prevent the Nova Compute service from starting, it should first be validated as described in the documentation <admin/managing-resource-providers>. The Nova Compute service can then be reconfigured to apply the change.

To remove the resource provider configuration, it is simplest to leave the group/host vars in place without specifying any inventory or traits. This will effectively remove the configuration when the Nova Compute service is restarted. If you choose to undefine nova_cell_compute_provider_config on a host, you must manually remove the generated config from inside the container, or recreate the container.