invoke-ai/InvokeAI · error · RuntimeError

Failed to load and migrate v3 config file {config_path}: {e}

Error message

Failed to load and migrate v3 config file {config_path}: {e}

What it means

load_and_migrate_config upgrades a v3 config file: it validates the loaded dict with DefaultInvokeAIAppConfig, writes the migrated file, and on any exception restores the .yaml.bak backup and re-raises as RuntimeError 'Failed to load and migrate v3 config file {path}: {e}'. This keeps the original config intact while signaling that migration failed.

Source

Thrown at invokeai/app/services/config/config_default.py:626

    if loaded_config_dict["schema_version"] == "4.0.0":
        migrated = True
        loaded_config_dict = migrate_v4_0_0_to_4_0_1_config_dict(loaded_config_dict)
    if loaded_config_dict["schema_version"] == "4.0.1":
        migrated = True
        loaded_config_dict = migrate_v4_0_1_to_4_0_2_config_dict(loaded_config_dict)
    if loaded_config_dict["schema_version"] == "4.0.2":
        migrated = True
        loaded_config_dict = migrate_v4_0_2_to_4_0_3_config_dict(loaded_config_dict)

    if migrated:
        shutil.copy(config_path, config_path.with_suffix(".yaml.bak"))
        try:
            # load and write without environment variables
            migrated_config = DefaultInvokeAIAppConfig.model_validate(loaded_config_dict)
            migrated_config.write_file(config_path)
        except Exception as e:
            shutil.copy(config_path.with_suffix(".yaml.bak"), config_path)
            raise RuntimeError(f"Failed to load and migrate v3 config file {config_path}: {e}") from e

    try:
        # Meta is not included in the model fields, so we need to validate it separately
        config = InvokeAIAppConfig.model_validate(loaded_config_dict)
        assert config.schema_version == CONFIG_SCHEMA_VERSION, (
            f"Invalid schema version, expected {CONFIG_SCHEMA_VERSION}: {config.schema_version}"
        )
        return config
    except Exception as e:
        raise RuntimeError(f"Failed to load config file {config_path}: {e}") from e


def load_external_api_keys(api_keys_file_path: Path) -> dict[str, str]:
    """Load external provider config (API keys and base URLs) from a dedicated YAML file."""
    if not api_keys_file_path.exists():
        return {}

    with open(api_keys_file_path, "rt", encoding=locale.getpreferredencoding()) as file:

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Read the inner '{e}' message to find the exact validation failure and fix that key in the YAML
2. Restore the automatic backup (config_path.yaml.bak was copied back) and fix types/keys before retrying
3. Rename or move the old config and let InvokeAI generate a fresh one, then re-add settings incrementally
4. Ensure the config directory is writable so write_file can persist the migrated config

Example fix

// before (invokeai.yaml, v3)
generation_devices: cuda:0   # invalid type for new schema
// after
generation_devices:
  - cuda:0

import yaml
data = yaml.safe_load(open('invokeai.yaml'))
from invokeai.app.services.config.config_default import DefaultInvokeAIAppConfig
DefaultInvokeAIAppConfig.model_validate(data)  # surfaces errors before migration writes

try:
    config = load_and_migrate_config(path)
except RuntimeError as e:
    logger.error(f"Config migration failed: {e}; .yaml.bak was restored")
    # fix the offending key reported in the chained exception, then retry

When it happens

Trigger: Running InvokeAI (or update_runtime_config/get_config) with a v3-era invokeai.yaml that fails model_validate — unknown/invalid keys, wrong types, malformed YAML values, or a write_file failure (permissions, read-only volume) — triggering the except branch.

Common situations: Upgrading InvokeAI across major versions with an old hand-edited config; config entries typed incorrectly (string where list expected, e.g. generation_devices); migrating configs in containers with read-only config mounts.

Related errors

• JWT secret not found in database. This should have been crea
• Invalid regex: {e}
• Invalid generation_devices value '{v}'. Use 'auto' or a list
• generation_devices cannot be an empty list. Use 'auto' or a
• base_url must not start with reserved path segment '/{first_

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/ad22808ca7bbf47d. Report an issue: GitHub.