Implement the normalisation and comparison of openshift config items

.gitlab-ci.yml

@@ -10,23 +10,29 @@ variables:
   PY_VERSION: "3.6"  # This is what we have running in OpenShift currently.
 
 
-## A full installation of CARA, tested with pytest.
-#test_install:
-#  extends: .acc_py_full_test
-#
-#
-## A development installation of CARA tested with pytest.
-#test_dev:
-#  extends: .acc_py_dev_test
+# A full installation of CARA, tested with pytest.
+test_install:
+  extends: .acc_py_full_test
 
 
 # A development installation of CARA tested with pytest.
-test_openshift_config:
+test_dev:
+  extends: .acc_py_dev_test
+
+
+# A development installation of CARA tested with pytest.
+.test_openshift_config:
+  variables:
+    CARA_INSTANCE: 'test-cara'
+    BRANCH: 'live/test-cara'
   rules:
+    - if: '$OPENSHIFT_CONFIG_CHECKER_TOKEN_TEST_CARA && $CI_MERGE_REQUEST_SOURCE_BRANCH_NAME == ${BRANCH}'
+      allow_failure: false  # The live/test-cara must represent what is deployed.
     - if: '$OPENSHIFT_CONFIG_CHECKER_TOKEN_TEST_CARA'
+      allow_failure: true   # Anything other than live/test-cara can fail.
   image: registry.cern.ch/docker.io/mambaorg/micromamba
   before_script:
-    - micromamba create --yes -p $HOME/env python=3.9 wget -c conda-forge
+    - micromamba create --yes -p $HOME/env python=3.9 ruamel.yaml wget -c conda-forge
     - export PATH=$HOME/env/bin/:$PATH
     - wget https://github.com/openshift/origin/releases/download/v3.11.0/openshift-origin-client-tools-v3.11.0-0cbc58b-linux-64bit.tar.gz
     - tar xzf ./openshift-origin-client-tools-v3.11.0-0cbc58b-linux-64bit.tar.gz
@@ -35,10 +41,12 @@ test_openshift_config:
   script:
     - cd ./app-config/openshift
     - oc login openshift-dev.cern.ch --token="${OPENSHIFT_CONFIG_CHECKER_TOKEN_TEST_CARA}"
-    - python ./fetch-config.py test-cara --output-directory ./test-cara/actual
-    - python ./generate-config.py test-cara --output-directory ./test-cara/expected
+    - python ./config-fetch.py test-cara --output-directory ./test-cara/actual
+    - python ./config-generate.py test-cara --output-directory ./test-cara/expected
+    - python ./config-normalise.py ./test-cara/actual ./test-cara/actual-normed
+    - python ./config-normalise.py ./test-cara/expected ./test-cara/expected-normed
+    - diff -u ./test-cara/actual-normed/ ./test-cara/expected-normed/
 
-#    - pytest ./test_config.py --arg test-cara
   artifacts:
     paths:
       - ./app-config/openshift/test-cara/actual

README.md

@@ -194,10 +194,12 @@ If you need to create the application in a new project, run:
 ```console
 $ cd app-config/openshift
 
-$ oc process -f application.yaml --param PROJECT_NAME='test-cara' --param GIT_BRANCH='live/test-cara' | oc create -f -
+$ oc process -f routes.yaml --param HOST='test-cara.web.cern.ch' | oc create -f -
 $ oc process -f configmap.yaml | oc create -f -
 $ oc process -f services.yaml | oc create -f -
-$ oc process -f route.yaml --param HOST='test-cara.web.cern.ch' | oc create -f -
+$ oc process -f imagestreams.yaml | oc create -f -
+$ oc process -f buildconfig.yaml --param GIT_BRANCH='live/test-cara' | oc create -f -
+$ oc process -f deploymentconfig.yaml --param PROJECT_NAME='test-cara'  | oc create -f -
 ```
 
 Then, create the webhook secret to be able to trigger automatic builds from GitLab.
@@ -272,7 +274,7 @@ $ cd app-config/openshift
 
 $ oc process -f configmap.yaml | oc replace -f -
 $ oc process -f services.yaml | oc replace -f -
-$ oc process -f route.yaml --param HOST='test-cara.web.cern.ch' | oc replace -f -
+$ oc process -f routes.yaml --param HOST='test-cara.web.cern.ch' | oc replace -f -
 $ oc process -f imagestreams.yaml | oc replace -f -
 $ oc process -f buildconfig.yaml --param GIT_BRANCH='live/test-cara' | oc replace -f -
 $ oc process -f deploymentconfig.yaml --param PROJECT_NAME='test-cara' | oc replace -f -
@@ -280,4 +282,3 @@ $ oc process -f deploymentconfig.yaml --param PROJECT_NAME='test-cara' | oc repl
 
 Be aware that if you change/replace the **route** of the PROD instance,
 it will lose the annotation to be exposed outside CERN (not committed in this repo).
-

app-config/openshift/config-fetch.py

@@ -13,7 +13,7 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
         help="Pick the instance for which you want to fetch the config",
     )
     parser.add_argument(
-        "--output-directory", default='config',
+        "-o", "--output-directory", default='config',
         help="Location to put the config files",
     )
 
@@ -36,8 +36,8 @@ def fetch_config(output_directory: pathlib.Path):
     output_directory.mkdir(exist_ok=True, parents=True)
 
     for component in ['routes', 'configmap', 'services', 'imagestreams', 'buildconfig', 'deploymentconfig']:
-        with (output_directory / f'{component}.json').open('wt') as fh:
-            cmd = ['oc', 'get', '--export', '-o', 'json', component]
+        with (output_directory / f'{component}.yaml').open('wt') as fh:
+            cmd = ['oc', 'get', '--export', '-o', 'yaml', component]
             print(f'Running: {" ".join(cmd)}')
             subprocess.run(cmd, stdout=fh, check=True)
     print(f'Config in: {output_directory.absolute()}')
@@ -53,7 +53,6 @@ def handler(args: argparse.ArgumentParser) -> None:
 
     actual_login_server = get_oc_server()
     if actual_login_server != login_server:
-        print(actual_login_server)
         print(f'\nPlease login to the correct openshift server with: \n\n oc login {login_server}\n', file=sys.stderr)
         sys.exit(1)
 

app-config/openshift/config-generate.py

@@ -1,7 +1,6 @@
 import argparse
 import pathlib
 import subprocess
-import sys
 import typing
 
 
@@ -13,7 +12,7 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
         help="Pick the instance for which you want to generate the config",
     )
     parser.add_argument(
-        "--output-directory", default='config',
+        "-o", "--output-directory", default='config',
         help="Location to put the config files",
     )
 
@@ -22,14 +21,13 @@ def generate_config(output_directory: pathlib.Path, project_name: str, hostname:
     output_directory.mkdir(exist_ok=True, parents=True)
 
     def oc_process(component_name: str, context: typing.Optional[dict] = None):
-        cmd = ['oc', 'process', '--local', '-f', f'{component_name}.yaml']
+        cmd = ['oc', 'process', '--local', '-f', f'{component_name}.yaml', '-o', 'yaml']
         for ctx_name, ctx_value in (context or {}).items():
             cmd.extend(['--param', f'{ctx_name}={ctx_value}'])
-        with (output_directory / f'{component_name}.json').open('wt') as fh:
+        with (output_directory / f'{component_name}.yaml').open('wt') as fh:
             print(f'Running: {" ".join(cmd)}')
             subprocess.run(cmd, stdout=fh, check=True)
 
-    # oc_process('route', oc_process + ['route.yaml', '--param', f'HOST={hostname}'])
     oc_process('routes', context={'HOST': hostname})
     oc_process('configmap')
     oc_process('services')

app-config/openshift/config-normalise.py

@@ -0,0 +1,117 @@
+import argparse
+import pathlib
+
+import ruamel.yaml
+
+
+def configure_parser(parser: argparse.ArgumentParser) -> None:
+    parser.description = "Normalise openshift config files (by sorting and removing ephemeral values)"
+    parser.set_defaults(handler=handler)
+    parser.add_argument(
+        "config-directory",
+        help="The directory from which to find yaml files",
+    )
+    parser.add_argument(
+        "output-directory",
+        help="The directory to put normalized files (can be the same as config-directory)",
+    )
+
+
+def clean_ephemeral_config(config: dict):
+    config = config.copy()
+    config.get('metadata', []).clear()
+
+    METADATA_TO_PRESERVE = ['labels', 'name']
+
+    for item in config['items']:
+        item.pop('status', None)
+
+        for key in list(item['metadata'].keys()):
+            if key not in METADATA_TO_PRESERVE:
+                del item['metadata'][key]
+
+        item.get('spec', {}).pop('clusterIP', None)
+
+        if item['kind'] == 'BuildConfig':
+            for trigger in item.get('spec', {}).get('triggers', []):
+                trigger.get('imageChange', {}).pop('lastTriggeredImageID', None)
+
+        if item['kind'] == 'DeploymentConfig':
+            item['spec'].get('template', {}).get('metadata', {}).pop('creationTimestamp', None)
+
+            for container in item['spec'].get('template', {}).get('spec', {}).get('containers', []):
+                # Drop the specific image name (and hash).
+                container.pop('image', None)
+            item['spec'].get('template', {}).get('metadata', {}).pop('creationTimestamp', None)
+            for trigger in item['spec'].get('triggers', []):
+                trigger.get('imageChangeParams', {}).pop('lastTriggeredImage', None)
+
+        # Drop the template part of the config for now.
+        # TODO: Remove this constraint to ensure our deployments reflect the fact that they are templated.
+        r = item['metadata'].get('labels', {}).pop('template', None)
+
+        if r is not None and not item['metadata']['labels']:
+            # Remove the empty labels dict if there is nothing left after popping the template item.
+            item['metadata'].pop('labels')
+
+    return config
+
+
+def deep_sort(item):
+    if isinstance(item, dict):
+        # Sort by the key.
+        return {k: deep_sort(v) for k, v in sorted(item.items(), key=lambda i: i[0])}
+    elif isinstance(item, list):
+        # Use the metadata/name and fallback to the str representation to give a sort order.
+        def sort_key(value):
+            if isinstance(value, dict):
+                return value.get('metadata', {}).get('name', '') or str(value)
+            else:
+                return str(value)
+
+        return sorted(
+            [deep_sort(v) for v in item],
+            key=sort_key,
+        )
+    else:
+        return item
+
+
+def normalise_config(input_directory: pathlib.Path, output_directory: pathlib.Path):
+    output_directory.mkdir(exist_ok=True, parents=True)
+
+    files = sorted(input_directory.glob('*.yaml'))
+
+    yaml = ruamel.yaml.YAML(typ='safe')
+
+    for file in files:
+        with file.open('rt') as fh:
+            content = yaml.load(fh)
+
+        config = clean_ephemeral_config(content)
+        config = deep_sort(config)
+
+        destination = output_directory / file.name
+        with destination.open('wt') as fh:
+            yaml.dump(config, fh)
+        print(f'Normalised {file.name} in {destination}')
+    print(f'Config in: {output_directory.absolute()}')
+
+
+def handler(args: argparse.ArgumentParser) -> None:
+
+    normalise_config(
+        pathlib.Path(getattr(args, 'config-directory')),
+        pathlib.Path(getattr(args, 'output-directory')),
+    )
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    configure_parser(parser)
+    args = parser.parse_args()
+    args.handler(args)
+
+
+if __name__ == '__main__':
+    main()