OpenStack: fix consumption of vendor-data to allow namespacing

Not all vendor data is destined for cloud-init.  This sanely reads
the vendor data as a dict, array or a string.

4 files changed, 77 insertions, 14 deletions

diff --git a/cloudinit/sources/DataSourceConfigDrive.py b/cloudinit/sources/DataSourceConfigDrive.py
index a27d07fb..4e5d90de 100644
--- a/cloudinit/sources/DataSourceConfigDrive.py
+++ b/cloudinit/sources/DataSourceConfigDrive.py
@@ -126,12 +126,13 @@ class DataSourceConfigDrive(openstack.SourceMixin, sources.DataSource):
         self.version = results['version']
         self.files.update(results.get('files', {}))
 
-        # If there is no vendordata, set vd to an empty dict instead of None
-        vd = results.get('vendordata', {})
-        # if vendordata includes 'cloud-init', then read that explicitly
-        # for cloud-init (for namespacing).
-        if 'cloud-init' in vd:
-            self.vendordata_raw = vd['cloud-init']
+        vd = results.get('vendordata')
+        self.vendordata_pure = vd
+        try:
+            self.vendordata_raw = openstack.convert_vendordata_json(vd)
+        except ValueError as e:
+            LOG.warn("Invalid content in vendor-data: %s", e)
+            self.vendordata_raw = None
 
         return True
 
diff --git a/cloudinit/sources/DataSourceOpenStack.py b/cloudinit/sources/DataSourceOpenStack.py
index 765137c6..469c2e2a 100644
--- a/cloudinit/sources/DataSourceOpenStack.py
+++ b/cloudinit/sources/DataSourceOpenStack.py
@@ -140,13 +140,13 @@ class DataSourceOpenStack(openstack.SourceMixin, sources.DataSource):
         self.version = results['version']
         self.files.update(results.get('files', {}))
 
-        # if vendordata includes 'cloud-init', then read that explicitly
-        # for cloud-init (for namespacing).
         vd = results.get('vendordata')
-        if isinstance(vd, dict) and 'cloud-init' in vd:
-            self.vendordata_raw = vd['cloud-init']
-        else:
-            self.vendordata_raw = vd
+        self.vendordata_pure = vd
+        try:
+            self.vendordata_raw = openstack.convert_vendordata_json(vd)
+        except ValueError as e:
+            LOG.warn("Invalid content in vendor-data: %s", e)
+            self.vendordata_raw = None
 
         return True
 
diff --git a/cloudinit/sources/helpers/openstack.py b/cloudinit/sources/helpers/openstack.py
index a7dd05df..b7e19314 100644
--- a/cloudinit/sources/helpers/openstack.py
+++ b/cloudinit/sources/helpers/openstack.py
@@ -21,6 +21,7 @@
 import abc
 import base64
 import copy
+import functools
 import os
 
 from cloudinit import ec2_utils
@@ -203,6 +204,9 @@ class BaseReader(object):
         If not a valid location, raise a NonReadable exception.
         """
 
+        load_json_anytype = functools.partial(
+            util.load_json, root_types=(dict, basestring, list))
+
         def datafiles(version):
             files = {}
             files['metadata'] = (
@@ -221,7 +225,7 @@ class BaseReader(object):
             files['vendordata'] = (
                 self._path_join("openstack", version, 'vendor_data.json'),
                 False,
-                util.load_json,
+                load_json_anytype,
             )
             return files
 
@@ -459,3 +463,28 @@ class MetadataReader(BaseReader):
         return ec2_utils.get_instance_metadata(ssl_details=self.ssl_details,
                                                timeout=self.timeout,
                                                retries=self.retries)
+
+
+def convert_vendordata_json(data, recurse=True):
+    """ data: a loaded json *object* (strings, arrays, dicts).
+    return something suitable for cloudinit vendordata_raw.
+
+    if data is:
+       None: return None
+       string: return string
+       list: return data
+             the list is then processed in UserDataProcessor
+       dict: return convert_vendordata_json(data.get('cloud-init'))
+    """
+    if not data:
+        return None
+    if isinstance(data, (str, unicode, basestring)):
+        return data
+    if isinstance(data, list):
+        return copy.deepcopy(data)
+    if isinstance(data, dict):
+        if recurse is True:
+            return convert_vendordata_json(data.get('cloud-init'),
+                                           recurse=False)
+        raise ValueError("vendordata['cloud-init'] cannot be dict")
+    raise ValueError("Unknown data type for vendordata: %s" % type(data))
diff --git a/tests/unittests/test_datasource/test_openstack.py b/tests/unittests/test_datasource/test_openstack.py
index 6809823e..7b4e651a 100644
--- a/tests/unittests/test_datasource/test_openstack.py
+++ b/tests/unittests/test_datasource/test_openstack.py
@@ -19,6 +19,7 @@
 import copy
 import json
 import re
+import unittest
 
 from StringIO import StringIO
 
@@ -256,7 +257,8 @@ class TestOpenStackDataSource(test_helpers.HttprettyTestCase):
         self.assertEquals(EC2_META, ds_os.ec2_metadata)
         self.assertEquals(USER_DATA, ds_os.userdata_raw)
         self.assertEquals(2, len(ds_os.files))
-        self.assertEquals(VENDOR_DATA, ds_os.vendordata_raw)
+        self.assertEquals(VENDOR_DATA, ds_os.vendordata_pure)
+        self.assertEquals(ds_os.vendordata_raw, None)
 
     @hp.activate
     def test_bad_datasource_meta(self):
@@ -314,3 +316,34 @@ class TestOpenStackDataSource(test_helpers.HttprettyTestCase):
         found = ds_os.get_data()
         self.assertFalse(found)
         self.assertIsNone(ds_os.version)
+
+
+class TestVendorDataLoading(unittest.TestCase):
+    def cvj(self, data):
+        return openstack.convert_vendordata_json(data)
+
+    def test_vd_load_none(self):
+        # non-existant vendor-data should return none
+        self.assertIsNone(self.cvj(None))
+
+    def test_vd_load_string(self):
+        self.assertEqual(self.cvj("foobar"), "foobar")
+
+    def test_vd_load_list(self):
+        data = [{'foo': 'bar'}, 'mystring', list(['another', 'list'])]
+        self.assertEqual(self.cvj(data), data)
+
+    def test_vd_load_dict_no_ci(self):
+        self.assertEqual(self.cvj({'foo': 'bar'}), None)
+
+    def test_vd_load_dict_ci_dict(self):
+        self.assertRaises(ValueError, self.cvj,
+                          {'foo': 'bar', 'cloud-init': {'x': 1}})
+
+    def test_vd_load_dict_ci_string(self):
+        data = {'foo': 'bar', 'cloud-init': 'VENDOR_DATA'}
+        self.assertEqual(self.cvj(data), data['cloud-init'])
+
+    def test_vd_load_dict_ci_list(self):
+        data = {'foo': 'bar', 'cloud-init': ['VD_1', 'VD_2']}
+        self.assertEqual(self.cvj(data), data['cloud-init'])