[go: up one dir, main page]

Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Define a container functionality to organize information to define a mock server. #674

Open
KSchopmeyer opened this issue Jul 16, 2020 · 9 comments

Comments

@KSchopmeyer
Copy link
Contributor
KSchopmeyer commented Jul 16, 2020

The mock server definition is becomming more and more complex. We need to define a container so that the various components of a mock server can be contained in a single file. This would include data like:

  1. The mof files that make up the mock and any install pywbem files required to build the repository.

  2. The schema used if one is required. This would include the schema source, version, and the leaf classes required.

  3. The namespaces required

  4. Basic serve information like server name, objectmanager name, etc.

  5. The user-defined providers required.

  6. Registered Profiles required

  7. Packaged profile definitions to install. This may be part of 2 above also, at least to install the classes.

Pywbemcli would analyze this file on startup and create the repository.

@KSchopmeyer KSchopmeyer added this to the 0.8.0 milestone Jul 16, 2020
@KSchopmeyer
Copy link
Contributor Author
KSchopmeyer commented Jul 16, 2020

We did this for the test environment using a dictionary. the form of that dictionary is as follows:

DEFAULT_WBEM_SERVER_MOCK_DICT = {
    'dmtf_schema': {'version': DMTF_TEST_SCHEMA_VER,
                    'dir': TESTSUITE_SCHEMA_DIR},
    'class_names': ['CIM_Namespace',
                    'CIM_ObjectManager',
                    'CIM_RegisteredProfile',
                    'CIM_ElementConformsToProfile',
                    'CIM_ReferencedProfile',
                    'CIM_ComputerSystem'],
    'class-mof': ["class MCK_StorageComputerSystem : CIM_ComputerSystem{};", ],
    'system_name': 'Mock_WBEMServerTest',
    'object_manager': {'Name': 'FakeObjectManager',
                       'ElementName': 'Pegasus',
                       'Description': 'Pegasus CIM Server Version 2.15.0'
                                      ' Released', },
    'interop_namspace': 'interop',
    'other_namespaces': [],
    'registered_profiles': [('DMTF', 'Indications', '1.1.0'),
                            ('DMTF', 'Profile Registration', '1.0.0'),
                            ('SNIA', 'Server', '1.2.0'),
                            ('SNIA', 'Server', '1.1.0'),
                            ('SNIA', 'SMI-S', '1.2.0'),
                            ('SNIA', 'Array', '1.4.0'),
                            ('SNIA', 'Software', '1.4.0'),
                            ('DMTF', 'Component', '1.4.0'), ],

    'referenced_profiles': [
        (('SNIA', 'Server', '1.2.0'), ('DMTF', 'Indications', '1.1.0')),
        (('SNIA', 'Server', '1.2.0'), ('SNIA', 'Array', '1.4.0')),
        (('SNIA', 'Array', '1.4.0'), ('DMTF', 'Component', '1.4.0')),
        (('SNIA', 'Array', '1.4.0'), ('SNIA', 'Software', '1.4.0')),

@KSchopmeyer
Copy link
Contributor Author
KSchopmeyer commented Jul 17, 2020
  1. Either yaml file or dictionary. We agreed that this should be yaml.

  2. The idea of packaging each profile as a separate entity is logical.

  3. The packages we could consider separately are the mock_server itself including parameters, interop namespace, etc, profile including their names, and all classes/instances, and user-providers (are there user-providers outside of profiles???).

@KSchopmeyer
Copy link
Contributor Author
KSchopmeyer commented Jul 31, 2020

We agreed that yaml is a logical format, it should have essentially the same characteristics as a python dictionary so that just about everything we could do with a python dictionary as the controlling structure we could do with yaml.

The goal: To package the requirements for installation of a defined profile as a single file so that it could be installed by a package like pywbemcli. NOTE: It is not clear that this is just a pywbemcli requirement but it is a good place to start.

Packaging a profile requirements:

  1. Name and version of the profile.
  2. Prerequisite profiles and providers. What must be installed before this profile can be installed.
  3. Namespace for the profile installation. Since this installs the profile itself, the only connection to registrations is that a profile registration mechanism must have already been installed.. Thus a prerequisite would be a profile_registration profile. This means that since profile implementations are defined in a user namespace, the prerequisite is that the user namespace be defined.
  4. Classes that are required for the profile. Note that this can be any of several ways of defining the classes including:
    a. dmtf schema and leaf class list (
    b. MOF with classes ( MOF file(s)
    c. CIM Objects to be installed (Python)
    3, Instances of classes for the profile including all associations.

Actions required of the install process:

  1. Validate prerequisites. If we can get everything packaged then all that we would have to do is to validate previous mock component packages (ex. profiles) have been installed.

  2. Determine that the user namespace exists or create it.

  3. Install the required classes. Note that this is only the classes for this profile. Classes like RegisteredProfile, etc. are installed as part of their own package.

  4. Install the instances of the profile.

  5. Define the registration components (RegisteredProfile, Implements...). This should simply execute something defined by the package that defiles the profile registration components so that this should be a call to something else with just the scoping / central instances and the profile definition (name, org, version).

TODO: Expand this for component profiles.

@KSchopmeyer
Copy link
Contributor Author
KSchopmeyer commented Aug 7, 2020

Definition of the mock-server

  1. creates conn.
  2. defines interop namespace
  3. defines dmtf schema (optional)
  4. Defines user providers required outside of profiles.
  5. Defines other namespaces (Optional)
  6. Sets up any special requirements (verbosity statistics collection, etc.) that would be part of the mock.

RESULT: Conn created, interop namespace created, other namespaces created. installs any names user providers.

@KSchopmeyer
Copy link
Contributor Author
  1. Allow definition of schema at the level of the profile. So that each profile could install install from a separate DMTF schema.

  2. We are talking about an aggregator mechanism and references other components (mof files, etc.)

  3. This should go into pywbem_mock so that the packaging concept and the methods that use the yaml are in pywbem mock but available from tools like pywbemcli.

@andy-maier andy-maier modified the milestones: 0.8.0, 0.9.0 Sep 16, 2020
@KSchopmeyer
Copy link
Contributor Author

Move to 0.9.0. Give ourselves time to consider based on work with caching.

@andy-maier
Copy link
Contributor

DISCUSSION: Do we still need this given that we have mock scripts?

@andy-maier
Copy link
Contributor

We decided to keep this one open and karl wants to come up with a more concrete proposal.

@KSchopmeyer
Copy link
Contributor Author

We agreed to explore those things that are difficult to do in the current scripts and see what can be simplified with new or modified function definitions.

@andy-maier andy-maier modified the milestones: 0.9.0, 1.0.0 Apr 16, 2021
@andy-maier andy-maier modified the milestones: 1.0.0, 1.1.0 Jul 17, 2021
@andy-maier andy-maier removed this from the 1.1.0 milestone Mar 15, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants