xref: /openbmc/openbmc-test-automation/docs/redfish_coding_guidelines.md (revision c17b98729fa0f26a4b9bf0753c14772936007525)

# Redfish Coding Guidelines

- For robot programs wishing to run Redfish commands, include the following in
  your robot file:

  ```
  *** Settings ***

  Resource                    bmc_redfish_resource.robot
  ```

- This git repository has some redfish wrapper modules:

  - [redfish_plus.py](../lib/redfish_plus.py)
  - [bmc_redfish.py](../lib/bmc_redfish.py)
  - [bmc_redfish_utils.py](../lib/bmc_redfish_utils.py)
  - Redfish wrapper module features:

    For all Redfish REST requests (get, head, post, put, patch, delete):

    - Support for python-like strings for all arguments which allows callers to
      easily specify complex arguments such as lists or dictionaries.

      So instead of coding this:

      ```
          ${ldap_type_dict}=  Create Dictionary  ServiceEnabled=${False}
          ${body}=  Create Dictionary  ${LDAP_TYPE}=${ldap_type_dict}
          Redfish.Patch  ${REDFISH_BASE_URI}AccountService  body=${body}
      ```

      You can do it in one fell swoop like this:

      ```
          Redfish.Patch  ${REDFISH_BASE_URI}AccountService  body={'${LDAP_TYPE}': {'ServiceEnabled': ${False}}}
      ```

    - Support for **valid_status_codes** argument and auto-failure:

      As mentioned above, this argument may be either an actual robot/python
      list or it may be a string value which python can translate into a list.

      The default value is [${HTTP_OK}].

      This means that the Redfish REST request will fail **automatically** if
      the resulting status code is not found in the valid_status_codes list.

      So instead of having to do this:

      ```
          ${resp}=  Redfish.Get  ${EVENT_LOG_URI}Entries
          Should Be Equal As Strings  ${resp.status_code}  ${HTTP_OK}
      ```

      You can simply do this:

      ```
          ${resp}=  Redfish.Get  ${EVENT_LOG_URI}Entries
      ```

      If, for some reason, you **expect** your command to fail, you can specify
      the expected status code or codes:

      ```
      Redfish.Patch  ${REDFISH_BASE_URI}UpdateService  body={'ApplyTime' : 'Invalid'}  valid_status_codes=[${HTTP_BAD_REQUEST}]
      ```

  - Login defaults for path, username and password are
    https://${OPENBMC_HOST},
    ${OPENBMC_USERNAME}, ${OPENBMC_PASSWORD}.
  - Many utility functions are available. Examples:;

    - get_properties
    - get_attributes
    - get_session_info
    - list_request
    - enumerate_request

# Rules for use of Redfish.Login and Redfish.Logout

It is desirable to avoid excessive redfish logins/logouts for the following
reasons:

- It simplifies the code base.
- It allows calling keywords and testcases to keep control over login parameters
  like USERNAME, PASSWORD, etc. Consider the following example:

  ```
  # Login to redfish with non-standard username/password.
  Redfish.Login  ${LDAP_USER}  ${LDAP_USER_PASSWORD}
  # Run 'Some Keyword' while logged in as ${LDAP_USER}/${LDAP_USER_PASSWORD}.
  Some Keyword
  ```

  If 'Some Keyword' in the example above does its own Redfish.Login, it will
  thwart the stated purpose of the caller.

**Rules:**

- Login should be done once in Suite Setup:

  ```
  *** Keywords ***
  Suite Setup Execution
      Redfish.Login
  ```

- Logout should be done once in Suite Teardown:
  ```
  *** Keywords ***
  Suite Teardown Execution
      Redfish.Logout
  ```
- As a result of the first two rules, all keywords and testcases that call upon
  redfish functions (e.g. Redfish.Get, Redfish.Patch, etc.) have a right to
  expect that login/logout have already been handled. Therefore, such keywords
  and testcases should NOT do logins and logouts themselves.
- There may be exceptions to the above but they require justification (e.g. a
  test whose purpose is to verify that it can login with an **alternate**
  username, etc.).
- Any keyword or test case which breaks the above rules is responsible for
  setting things right (i.e. back to a logged in state).

# Rules for use of data/variables.py

Avoid defining variables in data/variables.py for Redfish URIs.

There's no obvious benefit to using such variables. Conversely, with literal
values, it is much easier for the programmer to interpret the code.

Consider the following example.

Here's an excerpt from data/variables.py:

```
# Redfish variables.
REDFISH_BASE_URI = '/redfish/v1/'
...
REDFISH_ACCOUNTS = 'AccountService/Accounts/'
REDFISH_ACCOUNTS_URI = REDFISH_BASE_URI + REDFISH_ACCOUNTS
```

And here is a corresponding Robot code example:

```
    # Rather than coding this:
    Redfish.Delete  ${REDFISH_ACCOUNTS_URI}user_user

    # Code this:
    Redfish.Delete  /redfish/v1/AccountService/Accounts/user_user
```