ntp: Add schema definition and passive schema validation.

cloud-config files are very flexible and permissive.
This adds a jsonsschema definition to the cc_ntp module and validation
functions in cloudinit/config/schema which will log warnings about
invalid configuration values in the ntp section.

A cmdline tools/cloudconfig-schema is added which can be used in our dev
environments to quickly attempt to exercise the ntp schema.
It is also exposed as a main in cloudinit.config.schema.
(python3 -m cloudinit.config.schema)

LP: #1692916

7 files changed, 657 insertions, 7 deletions

diff --git a/cloudinit/config/cc_ntp.py b/cloudinit/config/cc_ntp.py
index 5cc54536..31ed64e3 100644
--- a/cloudinit/config/cc_ntp.py
+++ b/cloudinit/config/cc_ntp.py
@@ -36,6 +36,7 @@ servers or pools are provided, 4 pools will be used in the format
             - 192.168.23.2
 """
 
+from cloudinit.config.schema import validate_cloudconfig_schema
 from cloudinit import log as logging
 from cloudinit.settings import PER_INSTANCE
 from cloudinit import templater
@@ -43,6 +44,7 @@ from cloudinit import type_utils
 from cloudinit import util
 
 import os
+from textwrap import dedent
 
 LOG = logging.getLogger(__name__)
 
@@ -52,21 +54,84 @@ NR_POOL_SERVERS = 4
 distros = ['centos', 'debian', 'fedora', 'opensuse', 'ubuntu']
 
 
+# The schema definition for each cloud-config module is a strict contract for
+# describing supported configuration parameters for each cloud-config section.
+# It allows cloud-config to validate and alert users to invalid or ignored
+# configuration options before actually attempting to deploy with said
+# configuration.
+
+schema = {
+    'id': 'cc_ntp',
+    'name': 'NTP',
+    'title': 'enable and configure ntp',
+    'description': dedent("""\
+        Handle ntp configuration. If ntp is not installed on the system and
+        ntp configuration is specified, ntp will be installed. If there is a
+        default ntp config file in the image or one is present in the
+        distro's ntp package, it will be copied to ``/etc/ntp.conf.dist``
+        before any changes are made. A list of ntp pools and ntp servers can
+        be provided under the ``ntp`` config key. If no ntp ``servers`` or
+        ``pools`` are provided, 4 pools will be used in the format
+        ``{0-3}.{distro}.pool.ntp.org``."""),
+    'distros': distros,
+    'examples': [
+        {'ntp': {'pools': ['0.company.pool.ntp.org', '1.company.pool.ntp.org',
+                           'ntp.myorg.org'],
+                 'servers': ['my.ntp.server.local', 'ntp.ubuntu.com',
+                             '192.168.23.2']}}],
+    'frequency': PER_INSTANCE,
+    'type': 'object',
+    'properties': {
+        'ntp': {
+            'type': ['object', 'null'],
+            'properties': {
+                'pools': {
+                    'type': 'array',
+                    'items': {
+                        'type': 'string',
+                        'format': 'hostname'
+                    },
+                    'uniqueItems': True,
+                    'description': dedent("""\
+                        List of ntp pools. If both pools and servers are
+                         empty, 4 default pool servers will be provided of
+                         the format ``{0-3}.{distro}.pool.ntp.org``.""")
+                },
+                'servers': {
+                    'type': 'array',
+                    'items': {
+                        'type': 'string',
+                        'format': 'hostname'
+                    },
+                    'uniqueItems': True,
+                    'description': dedent("""\
+                        List of ntp servers. If both pools and servers are
+                         empty, 4 default pool servers will be provided with
+                         the format ``{0-3}.{distro}.pool.ntp.org``.""")
+                }
+            },
+            'required': [],
+            'additionalProperties': False
+        }
+    }
+}
+
+
 def handle(name, cfg, cloud, log, _args):
     """Enable and configure ntp."""
-
     if 'ntp' not in cfg:
         LOG.debug(
             "Skipping module named %s, not present or disabled by cfg", name)
         return
-
     ntp_cfg = cfg.get('ntp', {})
 
+    # TODO drop this when validate_cloudconfig_schema is strict=True
     if not isinstance(ntp_cfg, (dict)):
         raise RuntimeError(("'ntp' key existed in config,"
                             " but not a dictionary type,"
                             " is a %s %instead"), type_utils.obj_name(ntp_cfg))
 
+    validate_cloudconfig_schema(cfg, schema)
     rename_ntp_conf()
     # ensure when ntp is installed it has a configuration file
     # to use instead of starting up with packaged defaults
diff --git a/cloudinit/config/schema.py b/cloudinit/config/schema.py
new file mode 100644
index 00000000..6400f005
--- /dev/null
+++ b/cloudinit/config/schema.py
@@ -0,0 +1,222 @@
+# This file is part of cloud-init. See LICENSE file for license information.
+"""schema.py: Set of module functions for processing cloud-config schema."""
+
+from __future__ import print_function
+
+from cloudinit.util import read_file_or_url
+
+import argparse
+import logging
+import os
+import sys
+import yaml
+
+SCHEMA_UNDEFINED = b'UNDEFINED'
+CLOUD_CONFIG_HEADER = b'#cloud-config'
+SCHEMA_DOC_TMPL = """
+{name}
+---
+**Summary:** {title}
+
+{description}
+
+**Internal name:** ``{id}``
+
+**Module frequency:** {frequency}
+
+**Supported distros:** {distros}
+
+**Config schema**:
+{property_doc}
+{examples}
+"""
+SCHEMA_PROPERTY_TMPL = '{prefix}**{prop_name}:** ({type}) {description}'
+
+
+class SchemaValidationError(ValueError):
+    """Raised when validating a cloud-config file against a schema."""
+
+    def __init__(self, schema_errors=()):
+        """Init the exception an n-tuple of schema errors.
+
+        @param schema_errors: An n-tuple of the format:
+            ((flat.config.key, msg),)
+        """
+        self.schema_errors = schema_errors
+        error_messages = [
+            '{0}: {1}'.format(config_key, message)
+            for config_key, message in schema_errors]
+        message = "Cloud config schema errors: {0}".format(
+            ', '.join(error_messages))
+        super(SchemaValidationError, self).__init__(message)
+
+
+def validate_cloudconfig_schema(config, schema, strict=False):
+    """Validate provided config meets the schema definition.
+
+    @param config: Dict of cloud configuration settings validated against
+        schema.
+    @param schema: jsonschema dict describing the supported schema definition
+       for the cloud config module (config.cc_*).
+    @param strict: Boolean, when True raise SchemaValidationErrors instead of
+       logging warnings.
+
+    @raises: SchemaValidationError when provided config does not validate
+        against the provided schema.
+    """
+    try:
+        from jsonschema import Draft4Validator, FormatChecker
+    except ImportError:
+        logging.warning(
+            'Ignoring schema validation. python-jsonschema is not present')
+        return
+    validator = Draft4Validator(schema, format_checker=FormatChecker())
+    errors = ()
+    for error in sorted(validator.iter_errors(config), key=lambda e: e.path):
+        path = '.'.join([str(p) for p in error.path])
+        errors += ((path, error.message),)
+    if errors:
+        if strict:
+            raise SchemaValidationError(errors)
+        else:
+            messages = ['{0}: {1}'.format(k, msg) for k, msg in errors]
+            logging.warning('Invalid config:\n%s', '\n'.join(messages))
+
+
+def validate_cloudconfig_file(config_path, schema):
+    """Validate cloudconfig file adheres to a specific jsonschema.
+
+    @param config_path: Path to the yaml cloud-config file to parse.
+    @param schema: Dict describing a valid jsonschema to validate against.
+
+    @raises SchemaValidationError containing any of schema_errors encountered.
+    @raises RuntimeError when config_path does not exist.
+    """
+    if not os.path.exists(config_path):
+        raise RuntimeError('Configfile {0} does not exist'.format(config_path))
+    content = read_file_or_url('file://{0}'.format(config_path)).contents
+    if not content.startswith(CLOUD_CONFIG_HEADER):
+        errors = (
+            ('header', 'File {0} needs to begin with "{1}"'.format(
+                config_path, CLOUD_CONFIG_HEADER.decode())),)
+        raise SchemaValidationError(errors)
+
+    try:
+        cloudconfig = yaml.safe_load(content)
+    except yaml.parser.ParserError as e:
+        errors = (
+            ('format', 'File {0} is not valid yaml. {1}'.format(
+                config_path, str(e))),)
+        raise SchemaValidationError(errors)
+    validate_cloudconfig_schema(
+        cloudconfig, schema, strict=True)
+
+
+def _get_property_type(property_dict):
+    """Return a string representing a property type from a given jsonschema."""
+    property_type = property_dict.get('type', SCHEMA_UNDEFINED)
+    if isinstance(property_type, list):
+        property_type = '/'.join(property_type)
+    item_type = property_dict.get('items', {}).get('type')
+    if item_type:
+        property_type = '{0} of {1}'.format(property_type, item_type)
+    return property_type
+
+
+def _get_property_doc(schema, prefix='    '):
+    """Return restructured text describing the supported schema properties."""
+    new_prefix = prefix + '    '
+    properties = []
+    for prop_key, prop_config in schema.get('properties', {}).items():
+        # Define prop_name and dscription for SCHEMA_PROPERTY_TMPL
+        description = prop_config.get('description', '')
+        properties.append(SCHEMA_PROPERTY_TMPL.format(
+            prefix=prefix,
+            prop_name=prop_key,
+            type=_get_property_type(prop_config),
+            description=description.replace('\n', '')))
+        if 'properties' in prop_config:
+            properties.append(
+                _get_property_doc(prop_config, prefix=new_prefix))
+    return '\n\n'.join(properties)
+
+
+def _get_schema_examples(schema, prefix=''):
+    """Return restructured text describing the schema examples if present."""
+    examples = schema.get('examples')
+    if not examples:
+        return ''
+    rst_content = '\n**Examples**::\n\n'
+    for example in examples:
+        example_yaml = yaml.dump(example, default_flow_style=False)
+        # Python2.6 is missing textwrapper.indent
+        lines = example_yaml.split('\n')
+        indented_lines = ['    {0}'.format(line) for line in lines]
+        rst_content += '\n'.join(indented_lines)
+    return rst_content
+
+
+def get_schema_doc(schema):
+    """Return reStructured text rendering the provided jsonschema.
+
+    @param schema: Dict of jsonschema to render.
+    @raise KeyError: If schema lacks an expected key.
+    """
+    schema['property_doc'] = _get_property_doc(schema)
+    schema['examples'] = _get_schema_examples(schema)
+    schema['distros'] = ', '.join(schema['distros'])
+    return SCHEMA_DOC_TMPL.format(**schema)
+
+
+def get_schema(section_key=None):
+    """Return a dict of jsonschema defined in any cc_* module.
+
+    @param: section_key: Optionally limit schema to a specific top-level key.
+    """
+    # TODO use util.find_modules in subsequent branch
+    from cloudinit.config.cc_ntp import schema
+    return schema
+
+
+def error(message):
+    print(message, file=sys.stderr)
+    return 1
+
+
+def get_parser():
+    """Return a parser for supported cmdline arguments."""
+    parser = argparse.ArgumentParser()
+    parser.add_argument('-c', '--config-file',
+                        help='Path of the cloud-config yaml file to validate')
+    parser.add_argument('-d', '--doc', action="store_true", default=False,
+                        help='Print schema documentation')
+    parser.add_argument('-k', '--key',
+                        help='Limit validation or docs to a section key')
+    return parser
+
+
+def main():
+    """Tool to validate schema of a cloud-config file or print schema docs."""
+    parser = get_parser()
+    args = parser.parse_args()
+    exclusive_args = [args.config_file, args.doc]
+    if not any(exclusive_args) or all(exclusive_args):
+        return error('Expected either --config-file argument or --doc')
+
+    schema = get_schema()
+    if args.config_file:
+        try:
+            validate_cloudconfig_file(args.config_file, schema)
+        except (SchemaValidationError, RuntimeError) as e:
+            return error(str(e))
+        print("Valid cloud-config file {0}".format(args.config_file))
+    if args.doc:
+        print(get_schema_doc(schema))
+    return 0
+
+
+if __name__ == '__main__':
+    sys.exit(main())
+
+
+# vi: ts=4 expandtab
diff --git a/requirements.txt b/requirements.txt
index 0c4951f5..60abab16 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -36,5 +36,8 @@ requests
 # For patching pieces of cloud-config together
 jsonpatch
 
+# For validating cloud-config sections per schema definitions
+jsonschema 
+
 # For Python 2/3 compatibility
 six
diff --git a/tests/unittests/helpers.py b/tests/unittests/helpers.py
index 9ff15993..e78abce2 100644
--- a/tests/unittests/helpers.py
+++ b/tests/unittests/helpers.py
@@ -19,10 +19,6 @@ try:
     from contextlib import ExitStack
 except ImportError:
     from contextlib2 import ExitStack
-try:
-    from cStringIO import StringIO
-except ImportError:
-    from io import StringIO
 
 from cloudinit import helpers as ch
 from cloudinit import util
@@ -102,7 +98,7 @@ class CiTestCase(TestCase):
         if self.with_logs:
             # Create a log handler so unit tests can search expected logs.
             logger = logging.getLogger()
-            self.logs = StringIO()
+            self.logs = six.StringIO()
             handler = logging.StreamHandler(self.logs)
             self.old_handlers = logger.handlers
             logger.handlers = [handler]
diff --git a/tests/unittests/test_handler/test_handler_ntp.py b/tests/unittests/test_handler/test_handler_ntp.py
index bc4277b7..6cafa63d 100644
--- a/tests/unittests/test_handler/test_handler_ntp.py
+++ b/tests/unittests/test_handler/test_handler_ntp.py
@@ -212,4 +212,113 @@ class TestNtp(FilesystemMockingTestCase):
             'Skipping module named cc_ntp, not present or disabled by cfg\n',
             self.logs.getvalue())
 
+    def test_ntp_handler_schema_validation_allows_empty_ntp_config(self):
+        """Ntp schema validation allows for an empty ntp: configuration."""
+        invalid_config = {'ntp': {}}
+        distro = 'ubuntu'
+        cc = self._get_cloud(distro)
+        ntp_conf = os.path.join(self.new_root, 'ntp.conf')
+        with open('{0}.tmpl'.format(ntp_conf), 'wb') as stream:
+            stream.write(NTP_TEMPLATE)
+        with mock.patch('cloudinit.config.cc_ntp.NTP_CONF', ntp_conf):
+            cc_ntp.handle('cc_ntp', invalid_config, cc, None, [])
+        self.assertNotIn('Invalid config:', self.logs.getvalue())
+        with open(ntp_conf) as stream:
+            content = stream.read()
+        default_pools = [
+            "{0}.{1}.pool.ntp.org".format(x, distro)
+            for x in range(0, cc_ntp.NR_POOL_SERVERS)]
+        self.assertEqual(
+            "servers []\npools {0}\n".format(default_pools),
+            content)
+
+    def test_ntp_handler_schema_validation_warns_non_string_item_type(self):
+        """Ntp schema validation warns of non-strings in pools or servers.
+
+        Schema validation is not strict, so ntp config is still be rendered.
+        """
+        invalid_config = {'ntp': {'pools': [123], 'servers': ['valid', None]}}
+        cc = self._get_cloud('ubuntu')
+        ntp_conf = os.path.join(self.new_root, 'ntp.conf')
+        with open('{0}.tmpl'.format(ntp_conf), 'wb') as stream:
+            stream.write(NTP_TEMPLATE)
+        with mock.patch('cloudinit.config.cc_ntp.NTP_CONF', ntp_conf):
+            cc_ntp.handle('cc_ntp', invalid_config, cc, None, [])
+        self.assertIn(
+            "Invalid config:\nntp.pools.0: 123 is not of type 'string'\n"
+            "ntp.servers.1: None is not of type 'string'",
+            self.logs.getvalue())
+        with open(ntp_conf) as stream:
+            content = stream.read()
+        self.assertEqual("servers ['valid', None]\npools [123]\n", content)
+
+    def test_ntp_handler_schema_validation_warns_of_non_array_type(self):
+        """Ntp schema validation warns of non-array pools or servers types.
+
+        Schema validation is not strict, so ntp config is still be rendered.
+        """
+        invalid_config = {'ntp': {'pools': 123, 'servers': 'non-array'}}
+        cc = self._get_cloud('ubuntu')
+        ntp_conf = os.path.join(self.new_root, 'ntp.conf')
+        with open('{0}.tmpl'.format(ntp_conf), 'wb') as stream:
+            stream.write(NTP_TEMPLATE)
+        with mock.patch('cloudinit.config.cc_ntp.NTP_CONF', ntp_conf):
+            cc_ntp.handle('cc_ntp', invalid_config, cc, None, [])
+        self.assertIn(
+            "Invalid config:\nntp.pools: 123 is not of type 'array'\n"
+            "ntp.servers: 'non-array' is not of type 'array'",
+            self.logs.getvalue())
+        with open(ntp_conf) as stream:
+            content = stream.read()
+        self.assertEqual("servers non-array\npools 123\n", content)
+
+    def test_ntp_handler_schema_validation_warns_invalid_key_present(self):
+        """Ntp schema validation warns of invalid keys present in ntp config.
+
+        Schema validation is not strict, so ntp config is still be rendered.
+        """
+        invalid_config = {
+            'ntp': {'invalidkey': 1, 'pools': ['0.mycompany.pool.ntp.org']}}
+        cc = self._get_cloud('ubuntu')
+        ntp_conf = os.path.join(self.new_root, 'ntp.conf')
+        with open('{0}.tmpl'.format(ntp_conf), 'wb') as stream:
+            stream.write(NTP_TEMPLATE)
+        with mock.patch('cloudinit.config.cc_ntp.NTP_CONF', ntp_conf):
+            cc_ntp.handle('cc_ntp', invalid_config, cc, None, [])
+        self.assertIn(
+            "Invalid config:\nntp: Additional properties are not allowed "
+            "('invalidkey' was unexpected)",
+            self.logs.getvalue())
+        with open(ntp_conf) as stream:
+            content = stream.read()
+        self.assertEqual(
+            "servers []\npools ['0.mycompany.pool.ntp.org']\n",
+            content)
+
+    def test_ntp_handler_schema_validation_warns_of_duplicates(self):
+        """Ntp schema validation warns of duplicates in servers or pools.
+
+        Schema validation is not strict, so ntp config is still be rendered.
+        """
+        invalid_config = {
+            'ntp': {'pools': ['0.mypool.org', '0.mypool.org'],
+                    'servers': ['10.0.0.1', '10.0.0.1']}}
+        cc = self._get_cloud('ubuntu')
+        ntp_conf = os.path.join(self.new_root, 'ntp.conf')
+        with open('{0}.tmpl'.format(ntp_conf), 'wb') as stream:
+            stream.write(NTP_TEMPLATE)
+        with mock.patch('cloudinit.config.cc_ntp.NTP_CONF', ntp_conf):
+            cc_ntp.handle('cc_ntp', invalid_config, cc, None, [])
+        self.assertIn(
+            "Invalid config:\nntp.pools: ['0.mypool.org', '0.mypool.org'] has "
+            "non-unique elements\nntp.servers: ['10.0.0.1', '10.0.0.1'] has "
+            "non-unique elements",
+            self.logs.getvalue())
+        with open(ntp_conf) as stream:
+            content = stream.read()
+        self.assertEqual(
+            "servers ['10.0.0.1', '10.0.0.1']\n"
+            "pools ['0.mypool.org', '0.mypool.org']\n",
+            content)
+
 # vi: ts=4 expandtab
diff --git a/tests/unittests/test_handler/test_schema.py b/tests/unittests/test_handler/test_schema.py
new file mode 100644
index 00000000..3239e326
--- /dev/null
+++ b/tests/unittests/test_handler/test_schema.py
@@ -0,0 +1,220 @@
+# This file is part of cloud-init. See LICENSE file for license information.
+
+from cloudinit.config.schema import (
+    CLOUD_CONFIG_HEADER, SchemaValidationError, get_schema_doc,
+    validate_cloudconfig_file, validate_cloudconfig_schema,
+    main)
+from cloudinit.util import write_file
+
+from ..helpers import CiTestCase, mock
+
+from copy import copy
+from six import StringIO
+from textwrap import dedent
+
+
+class SchemaValidationErrorTest(CiTestCase):
+    """Test validate_cloudconfig_schema"""
+
+    def test_schema_validation_error_expects_schema_errors(self):
+        """SchemaValidationError is initialized from schema_errors."""
+        errors = (('key.path', 'unexpected key "junk"'),
+                  ('key2.path', '"-123" is not a valid "hostname" format'))
+        exception = SchemaValidationError(schema_errors=errors)
+        self.assertIsInstance(exception, Exception)
+        self.assertEqual(exception.schema_errors, errors)
+        self.assertEqual(
+            'Cloud config schema errors: key.path: unexpected key "junk", '
+            'key2.path: "-123" is not a valid "hostname" format',
+            str(exception))
+        self.assertTrue(isinstance(exception, ValueError))
+
+
+class ValidateCloudConfigSchemaTest(CiTestCase):
+    """Tests for validate_cloudconfig_schema."""
+
+    with_logs = True
+
+    def test_validateconfig_schema_non_strict_emits_warnings(self):
+        """When strict is False validate_cloudconfig_schema emits warnings."""
+        schema = {'properties': {'p1': {'type': 'string'}}}
+        validate_cloudconfig_schema({'p1': -1}, schema, strict=False)
+        self.assertIn(
+            "Invalid config:\np1: -1 is not of type 'string'\n",
+            self.logs.getvalue())
+
+    def test_validateconfig_schema_emits_warning_on_missing_jsonschema(self):
+        """Warning from validate_cloudconfig_schema when missing jsonschema."""
+        schema = {'properties': {'p1': {'type': 'string'}}}
+        with mock.patch.dict('sys.modules', **{'jsonschema': ImportError()}):
+            validate_cloudconfig_schema({'p1': -1}, schema, strict=True)
+        self.assertIn(
+            'Ignoring schema validation. python-jsonschema is not present',
+            self.logs.getvalue())
+
+    def test_validateconfig_schema_strict_raises_errors(self):
+        """When strict is True validate_cloudconfig_schema raises errors."""
+        schema = {'properties': {'p1': {'type': 'string'}}}
+        with self.assertRaises(SchemaValidationError) as context_mgr:
+            validate_cloudconfig_schema({'p1': -1}, schema, strict=True)
+        self.assertEqual(
+            "Cloud config schema errors: p1: -1 is not of type 'string'",
+            str(context_mgr.exception))
+
+    def test_validateconfig_schema_honors_formats(self):
+        """When strict is True validate_cloudconfig_schema raises errors."""
+        schema = {
+            'properties': {'p1': {'type': 'string', 'format': 'hostname'}}}
+        with self.assertRaises(SchemaValidationError) as context_mgr:
+            validate_cloudconfig_schema({'p1': '-1'}, schema, strict=True)
+        self.assertEqual(
+            "Cloud config schema errors: p1: '-1' is not a 'hostname'",
+            str(context_mgr.exception))
+
+
+class ValidateCloudConfigFileTest(CiTestCase):
+    """Tests for validate_cloudconfig_file."""
+
+    def setUp(self):
+        super(ValidateCloudConfigFileTest, self).setUp()
+        self.config_file = self.tmp_path('cloudcfg.yaml')
+
+    def test_validateconfig_file_error_on_absent_file(self):
+        """On absent config_path, validate_cloudconfig_file errors."""
+        with self.assertRaises(RuntimeError) as context_mgr:
+            validate_cloudconfig_file('/not/here', {})
+        self.assertEqual(
+            'Configfile /not/here does not exist',
+            str(context_mgr.exception))
+
+    def test_validateconfig_file_error_on_invalid_header(self):
+        """On invalid header, validate_cloudconfig_file errors.
+
+        A SchemaValidationError is raised when the file doesn't begin with
+        CLOUD_CONFIG_HEADER.
+        """
+        write_file(self.config_file, '#junk')
+        with self.assertRaises(SchemaValidationError) as context_mgr:
+            validate_cloudconfig_file(self.config_file, {})
+        self.assertEqual(
+            'Cloud config schema errors: header: File {0} needs to begin with '
+            '"{1}"'.format(self.config_file, CLOUD_CONFIG_HEADER.decode()),
+            str(context_mgr.exception))
+
+    def test_validateconfig_file_error_on_non_yaml_format(self):
+        """On non-yaml format, validate_cloudconfig_file errors."""
+        write_file(self.config_file, '#cloud-config\n{}}')
+        with self.assertRaises(SchemaValidationError) as context_mgr:
+            validate_cloudconfig_file(self.config_file, {})
+        self.assertIn(
+            'schema errors: format: File {0} is not valid yaml.'.format(
+                self.config_file),
+            str(context_mgr.exception))
+
+    def test_validateconfig_file_sctricty_validates_schema(self):
+        """validate_cloudconfig_file raises errors on invalid schema."""
+        schema = {
+            'properties': {'p1': {'type': 'string', 'format': 'hostname'}}}
+        write_file(self.config_file, '#cloud-config\np1: "-1"')
+        with self.assertRaises(SchemaValidationError) as context_mgr:
+            validate_cloudconfig_file(self.config_file, schema)
+        self.assertEqual(
+            "Cloud config schema errors: p1: '-1' is not a 'hostname'",
+            str(context_mgr.exception))
+
+
+class GetSchemaDocTest(CiTestCase):
+    """Tests for get_schema_doc."""
+
+    def setUp(self):
+        super(GetSchemaDocTest, self).setUp()
+        self.required_schema = {
+            'title': 'title', 'description': 'description', 'id': 'id',
+            'name': 'name', 'frequency': 'frequency',
+            'distros': ['debian', 'rhel']}
+
+    def test_get_schema_doc_returns_restructured_text(self):
+        """get_schema_doc returns restructured text for a cloudinit schema."""
+        full_schema = copy(self.required_schema)
+        full_schema.update(
+            {'properties': {
+                'prop1': {'type': 'array', 'description': 'prop-description',
+                          'items': {'type': 'int'}}}})
+        self.assertEqual(
+            dedent("""
+                name
+                ---
+                **Summary:** title
+
+                description
+
+                **Internal name:** ``id``
+
+                **Module frequency:** frequency
+
+                **Supported distros:** debian, rhel
+
+                **Config schema**:
+                    **prop1:** (array of int) prop-description\n\n"""),
+            get_schema_doc(full_schema))
+
+    def test_get_schema_doc_returns_restructured_text_with_examples(self):
+        """get_schema_doc returns indented examples when present in schema."""
+        full_schema = copy(self.required_schema)
+        full_schema.update(
+            {'examples': {'ex1': [1, 2, 3]},
+             'properties': {
+                'prop1': {'type': 'array', 'description': 'prop-description',
+                          'items': {'type': 'int'}}}})
+        self.assertIn(
+            dedent("""
+                **Config schema**:
+                    **prop1:** (array of int) prop-description
+
+                **Examples**::
+
+                    ex1"""),
+            get_schema_doc(full_schema))
+
+    def test_get_schema_doc_raises_key_errors(self):
+        """get_schema_doc raises KeyErrors on missing keys."""
+        for key in self.required_schema:
+            invalid_schema = copy(self.required_schema)
+            invalid_schema.pop(key)
+            with self.assertRaises(KeyError) as context_mgr:
+                get_schema_doc(invalid_schema)
+            self.assertEqual("'{0}'".format(key), str(context_mgr.exception))
+
+
+class MainTest(CiTestCase):
+
+    def test_main_missing_args(self):
+        """Main exits non-zero and reports an error on missing parameters."""
+        with mock.patch('sys.argv', ['mycmd']):
+            with mock.patch('sys.stderr', new_callable=StringIO) as m_stderr:
+                self.assertEqual(1, main(), 'Expected non-zero exit code')
+        self.assertEqual(
+            'Expected either --config-file argument or --doc\n',
+            m_stderr.getvalue())
+
+    def test_main_prints_docs(self):
+        """When --doc parameter is provided, main generates documentation."""
+        myargs = ['mycmd', '--doc']
+        with mock.patch('sys.argv', myargs):
+            with mock.patch('sys.stdout', new_callable=StringIO) as m_stdout:
+                self.assertEqual(0, main(), 'Expected 0 exit code')
+        self.assertIn('\nNTP\n---\n', m_stdout.getvalue())
+
+    def test_main_validates_config_file(self):
+        """When --config-file parameter is provided, main validates schema."""
+        myyaml = self.tmp_path('my.yaml')
+        myargs = ['mycmd', '--config-file', myyaml]
+        with open(myyaml, 'wb') as stream:
+            stream.write(b'#cloud-config\nntp:')   # shortest ntp schema
+        with mock.patch('sys.argv', myargs):
+            with mock.patch('sys.stdout', new_callable=StringIO) as m_stdout:
+                self.assertEqual(0, main(), 'Expected 0 exit code')
+        self.assertIn(
+            'Valid cloud-config file {0}'.format(myyaml), m_stdout.getvalue())
+
+# vi: ts=4 expandtab syntax=python
diff --git a/tools/cloudconfig-schema b/tools/cloudconfig-schema
new file mode 100755
index 00000000..32f0d61e
--- /dev/null
+++ b/tools/cloudconfig-schema
@@ -0,0 +1,35 @@
+#!/usr/bin/env python3
+# This file is part of cloud-init. See LICENSE file for license information.
+
+"""cloudconfig-schema
+
+Validate existing files against cloud-config schema or provide supported schema
+documentation.
+"""
+
+import os
+import sys
+
+
+def call_entry_point(name):
+    (istr, dot, ent) = name.rpartition('.')
+    try:
+        __import__(istr)
+    except ImportError:
+        # if that import failed, check dirname(__file__/..)
+        # to support ./bin/program with modules in .
+        _tdir = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
+        sys.path.insert(0, _tdir)
+        try:
+            __import__(istr)
+        except ImportError as e:
+            sys.stderr.write("Unable to find %s: %s\n" % (name, e))
+            sys.exit(2)
+
+    sys.exit(getattr(sys.modules[istr], ent)())
+
+
+if __name__ == '__main__':
+    call_entry_point("cloudinit.config.schema.main")
+
+# vi: ts=4 expandtab syntax=python