NoCloudConfig

NoCloudConfig refers to the NoCloud datasource configuration used by cloud-init to provision machines in environments that lack a traditional cloud metadata service. It provides instance metadata and user data from local seeds rather than from a remote cloud provider, making it suitable for virtual machines, bare metal, or testing environments.

A NoCloud seed typically consists of at least two files: user-data and meta-data. The user-data file can contain cloud-init directives written in YAML, enabling tasks such as setting a hostname, creating users and SSH keys, installing packages, and running commands. The meta-data file can supply metadata items like instance-id and local-hostname. An optional vendor-data file may also be provided. A simple example of content might include a cloud-config header, a desired hostname, and an SSH key entry in the users section.

Seeds can be provided in several ways. They are usually delivered to the guest VM via a seed ISO image, a virtual disk mounted at boot, or a directory mounted into the VM containing the seed files. Cloud-init searches for a NoCloud seed in common paths such as /var/lib/cloud/seed/nocloud/ or /media/seed/nocloud/, or from an injected ISO, depending on the platform and configuration. When detected, cloud-init reads the user-data and meta-data on first boot to apply the requested configurations.

NoCloudConfig is widely used for provisioning in environments without dynamic metadata services. It is limited to static data provided at boot time and is not suitable for scenarios requiring runtime metadata updates. Alternatives include other cloud-init datasources such as OpenStack, EC2, or ConfigDrive, depending on the deployment environment.