How the vt_settings Plugin Manages VT-Specific Configuration Options in Avocado-VT

The vt_settings plugin injects VT-specific configuration files into Avocado's settings search path by prepending them to the configuration hierarchy, enabling seamless integration of virtualization testing defaults that users can override via standard Avocado configuration files.

The vt_settings plugin in the avocado-framework/avocado-vt repository provides a robust mechanism for managing virtualization testing (VT) configuration options. By implementing Avocado's Settings plugin interface, it ensures that VT-specific defaults from conf.d/*.conf files are loaded early in the configuration hierarchy, making them available to all Avocado-VT components while remaining fully overridable by end users.

Architecture of the vt_settings Plugin

The VTSettings Class and adjust_settings_paths()

The core implementation resides in avocado_vt/plugins/vt_settings.py, where the VTSettings class implements the Settings plugin interface from avocado.core.plugin_interfaces. The critical method is adjust_settings_paths(), which receives the mutable paths list during Avocado startup and modifies it to include VT-specific configuration files.

Plugin Registration via Entry Points

The plugin is registered in setup.py under the avocado.plugins.settings entry point group. This registration uses the identifier vt-settings mapped to avocado_vt.plugins.vt_settings:VTSettings, ensuring automatic discovery during Avocado's plugin initialization phase.

How VT Configuration Files Are Injected

During Avocado startup, the vt_settings plugin executes a specific sequence to insert VT defaults at the highest priority level. The adjust_settings_paths() method uses importlib.resources to locate the package-installed configuration directory:

def adjust_settings_paths(self, paths):
    base = importlib.resources.files("avocado_vt").joinpath("conf.d")
    for path in base.iterdir():
        if path.is_file() and path.name.endswith(".conf"):
            paths.insert(0, str(path))  # Prepend to highest priority

This code iterates through avocado_vt/conf.d/, identifying files like vt.conf and vt_joblock.conf. By inserting each file at index 0 of the paths list, the plugin ensures these VT-specific defaults are processed before any user-supplied configuration files. This ordering follows Avocado's configuration hierarchy where earlier entries take precedence during the merge process.

Practical Usage and Override Behavior

Once injected, VT configuration options behave like standard Avocado settings. You can view the merged configuration using the command line:

# Display the complete configuration including VT defaults

avocado config

# Filter for specific VT-related settings

avocado config | grep vt

Users override VT defaults by creating or editing ~/.config/avocado/avocado.conf. For example, to change the maximum parallel tasks from the VT default:

echo "[run]
max_parallel_tasks = 4" >> ~/.config/avocado/avocado.conf

When Avocado reloads, the user-defined value takes precedence over the VT default because the user configuration file appears later in the merged paths list.

Within test code, access these settings through Avocado's settings API:

from avocado.core import settings

# Retrieve current VT-related configuration

config = settings.as_dict()
max_tasks = config.get('run.max_parallel_tasks')
print(f"Configured max parallel tasks: {max_tasks}")

Because the plugin pre-loads vt.conf, the above call will return the default value unless the user has overridden it in their personal configuration.

Summary

• The vt_settings plugin implements Avocado's Settings interface to inject VT-specific configuration files into the startup sequence.
• Located in avocado_vt/plugins/vt_settings.py, the VTSettings class prepends .conf files from avocado_vt/conf.d/ to the configuration paths list.
• This prepending strategy ensures VT defaults load with highest priority while remaining fully overridable by user configuration files.
• Configuration files like vt.conf and vt_joblock.conf provide defaults for sections such as [vt], [vt.qemu], and [plugins.vtjoblock].
• Users interact with these settings through standard Avocado commands and APIs, with overrides stored in ~/.config/avocado/avocado.conf.

Frequently Asked Questions

What is the vt_settings plugin in Avocado-VT?

The vt_settings plugin is an Avocado settings extension that automatically injects virtualization testing (VT) configuration defaults into Avocado's configuration hierarchy. It ensures that VT-specific options from files like vt.conf are available to all Avocado-VT components without requiring manual configuration file management by users.

How does the vt_settings plugin prioritize configuration files?

The plugin prioritizes VT defaults by prepending them to the configuration paths list during Avocado startup. In the adjust_settings_paths() method, each .conf file from avocado_vt/conf.d/ is inserted at index 0 of the paths list, ensuring these files are processed first. Since Avocado merges configurations in path order, with earlier files taking precedence, this gives VT defaults the highest priority while still allowing user overrides from later files.

Can users override VT-specific configuration options?

Yes, users can override any VT-specific option by creating or editing their user configuration file at ~/.config/avocado/avocado.conf. Because the vt_settings plugin inserts VT defaults at the beginning of the configuration paths list, user configuration files loaded later in the sequence will override any conflicting values. Users can verify current settings using the avocado config command.

Where are the VT configuration files located in the source repository?

The VT configuration files are located in the avocado_vt/conf.d/ directory within the avocado-vt repository. Key files include vt.conf, which contains defaults for general VT settings and QEMU-specific options, and vt_joblock.conf, which configures the VT job lock plugin. These files are packaged with the avocado-vt installation and located at runtime using importlib.resources.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}