[PR #8434/961767e2 backport][stable-9] Redfish: Add options to check the availability of the service (#8457)

Redfish: Add options to check the availability of the service (#8434)

* Redfish: Add options to check the availability of the service

Signed-off-by: Mike Raineri <michael.raineri@dell.com>

* Updates based on review feedback

Signed-off-by: Mike Raineri <michael.raineri@dell.com>

* Updated comment to reflect changed behavior

Signed-off-by: Mike Raineri <michael.raineri@dell.com>

* Added changelog fragments

Signed-off-by: Mike Raineri <michael.raineri@dell.com>

* Update changelogs/fragments/8051-Redfish-Wait-For-Service.yml

Co-authored-by: Felix Fontein <felix@fontein.de>

* Update plugins/modules/redfish_command.py

Co-authored-by: Felix Fontein <felix@fontein.de>

* Update plugins/modules/redfish_command.py

Co-authored-by: Felix Fontein <felix@fontein.de>

* Update plugins/modules/redfish_command.py

Co-authored-by: Felix Fontein <felix@fontein.de>

* Update plugins/modules/redfish_command.py

Co-authored-by: Felix Fontein <felix@fontein.de>

---------

Signed-off-by: Mike Raineri <michael.raineri@dell.com>
Co-authored-by: Felix Fontein <felix@fontein.de>
(cherry picked from commit 961767e2dd)

Co-authored-by: Mike Raineri <mraineri@gmail.com>

changelogs/fragments/8051-Redfish-Wait-For-Service.yml

@@ -0,0 +1,3 @@
+minor_changes:
+  - redfish_info - add command ``CheckAvailability`` to check if a service is accessible (https://github.com/ansible-collections/community.general/issues/8051, https://github.com/ansible-collections/community.general/pull/8434).
+  - redfish_command - add ``wait`` and ``wait_timeout`` options to allow a user to block a command until a service is accessible after performing the requested command (https://github.com/ansible-collections/community.general/issues/8051, https://github.com/ansible-collections/community.general/pull/8434).

plugins/module_utils/redfish_utils.py

@@ -11,6 +11,7 @@ import os
 import random
 import string
 import gzip
+import time
 from io import BytesIO
 from ansible.module_utils.urls import open_url
 from ansible.module_utils.common.text.converters import to_native
@@ -132,11 +133,13 @@ class RedfishUtils(object):
         return resp
 
     # The following functions are to send GET/POST/PATCH/DELETE requests
-    def get_request(self, uri, override_headers=None, allow_no_resp=False):
+    def get_request(self, uri, override_headers=None, allow_no_resp=False, timeout=None):
         req_headers = dict(GET_HEADERS)
         if override_headers:
             req_headers.update(override_headers)
         username, password, basic_auth = self._auth_params(req_headers)
+        if timeout is None:
+            timeout = self.timeout
         try:
             # Service root is an unauthenticated resource; remove credentials
             # in case the caller will be using sessions later.
@@ -146,7 +149,7 @@ class RedfishUtils(object):
                             url_username=username, url_password=password,
                             force_basic_auth=basic_auth, validate_certs=False,
                             follow_redirects='all',
-                            use_proxy=True, timeout=self.timeout)
+                            use_proxy=True, timeout=timeout)
             headers = dict((k.lower(), v) for (k, v) in resp.info().items())
             try:
                 if headers.get('content-encoding') == 'gzip' and LooseVersion(ansible_version) < LooseVersion('2.14'):
@@ -624,6 +627,24 @@ class RedfishUtils(object):
             allowable_values = default_values
         return allowable_values
 
+    def check_service_availability(self):
+        """
+        Checks if the service is accessible.
+
+        :return: dict containing the status of the service
+        """
+
+        # Get the service root
+        # Override the timeout since the service root is expected to be readily
+        # available.
+        service_root = self.get_request(self.root_uri + self.service_root, timeout=10)
+        if service_root['ret'] is False:
+            # Failed, either due to a timeout or HTTP error; not available
+            return {'ret': True, 'available': False}
+
+        # Successfully accessed the service root; available
+        return {'ret': True, 'available': True}
+
     def get_logs(self):
         log_svcs_uri_list = []
         list_of_logs = []
@@ -1083,11 +1104,12 @@ class RedfishUtils(object):
         return self.manage_power(command, self.systems_uri,
                                  '#ComputerSystem.Reset')
 
-    def manage_manager_power(self, command):
+    def manage_manager_power(self, command, wait=False, wait_timeout=120):
         return self.manage_power(command, self.manager_uri,
-                                 '#Manager.Reset')
+                                 '#Manager.Reset', wait, wait_timeout)
 
-    def manage_power(self, command, resource_uri, action_name):
+    def manage_power(self, command, resource_uri, action_name, wait=False,
+                     wait_timeout=120):
         key = "Actions"
         reset_type_values = ['On', 'ForceOff', 'GracefulShutdown',
                              'GracefulRestart', 'ForceRestart', 'Nmi',
@@ -1147,6 +1169,30 @@ class RedfishUtils(object):
         response = self.post_request(self.root_uri + action_uri, payload)
         if response['ret'] is False:
             return response
+
+        # If requested to wait for the service to be available again, block
+        # until it's ready
+        if wait:
+            elapsed_time = 0
+            start_time = time.time()
+            # Start with a large enough sleep.  Some services will process new
+            # requests while in the middle of shutting down, thus breaking out
+            # early.
+            time.sleep(30)
+
+            # Periodically check for the service's availability.
+            while elapsed_time <= wait_timeout:
+                status = self.check_service_availability()
+                if status['available']:
+                    # It's available; we're done
+                    break
+                time.sleep(5)
+                elapsed_time = time.time() - start_time
+
+            if elapsed_time > wait_timeout:
+                # Exhausted the wait timer; error
+                return {'ret': False, 'changed': True,
+                        'msg': 'The service did not become available after %d seconds' % wait_timeout}
         return {'ret': True, 'changed': True}
 
     def manager_reset_to_defaults(self, command):

plugins/modules/redfish_command.py

@@ -288,6 +288,20 @@ options:
     type: str
     choices: [ ResetAll, PreserveNetworkAndUsers, PreserveNetwork ]
     version_added: 8.6.0
+  wait:
+    required: false
+    description:
+      - Block until the service is ready again.
+    type: bool
+    default: false
+    version_added: 9.1.0
+  wait_timeout:
+    required: false
+    description:
+      - How long to block until the service is ready again before giving up.
+    type: int
+    default: 120
+    version_added: 9.1.0
 
 author:
   - "Jose Delarosa (@jose-delarosa)"
@@ -685,6 +699,16 @@ EXAMPLES = '''
       username: "{{ username }}"
       password: "{{ password }}"
 
+  - name: Restart manager power gracefully and wait for it to be available
+    community.general.redfish_command:
+      category: Manager
+      command: GracefulRestart
+      resource_id: BMC
+      baseuri: "{{ baseuri }}"
+      username: "{{ username }}"
+      password: "{{ password }}"
+      wait: True
+
   - name: Restart manager power gracefully
     community.general.redfish_command:
       category: Manager
@@ -841,7 +865,9 @@ def main():
             ),
             strip_etag_quotes=dict(type='bool', default=False),
             reset_to_defaults_mode=dict(choices=['ResetAll', 'PreserveNetworkAndUsers', 'PreserveNetwork']),
-            bios_attributes=dict(type="dict")
+            bios_attributes=dict(type="dict"),
+            wait=dict(type='bool', default=False),
+            wait_timeout=dict(type='int', default=120),
         ),
         required_together=[
             ('username', 'password'),
@@ -1016,7 +1042,7 @@ def main():
                 command = 'PowerGracefulRestart'
 
             if command.startswith('Power'):
-                result = rf_utils.manage_manager_power(command)
+                result = rf_utils.manage_manager_power(command, module.params['wait'], module.params['wait_timeout'])
             elif command == 'ClearLogs':
                 result = rf_utils.clear_logs()
             elif command == 'VirtualMediaInsert':

plugins/modules/redfish_info.py

@@ -359,6 +359,16 @@ EXAMPLES = '''
       baseuri: "{{ baseuri }}"
       username: "{{ username }}"
       password: "{{ password }}"
+
+  - name: Check the availability of the service with a timeout of 5 seconds
+    community.general.redfish_info:
+      category: Service
+      command: CheckAvailability
+      baseuri: "{{ baseuri }}"
+      username: "{{ username }}"
+      password: "{{ password }}"
+      timeout: 5
+    register: result
 '''
 
 RETURN = '''
@@ -385,6 +395,7 @@ CATEGORY_COMMANDS_ALL = {
                "GetUpdateStatus"],
     "Manager": ["GetManagerNicInventory", "GetVirtualMedia", "GetLogs", "GetNetworkProtocols",
                 "GetHealthReport", "GetHostInterfaces", "GetManagerInventory", "GetServiceIdentification"],
+    "Service": ["CheckAvailability"],
 }
 
 CATEGORY_COMMANDS_DEFAULT = {
@@ -393,7 +404,8 @@ CATEGORY_COMMANDS_DEFAULT = {
     "Accounts": "ListUsers",
     "Update": "GetFirmwareInventory",
     "Sessions": "GetSessions",
-    "Manager": "GetManagerNicInventory"
+    "Manager": "GetManagerNicInventory",
+    "Service": "CheckAvailability",
 }
 
 
@@ -473,7 +485,13 @@ def main():
             module.fail_json(msg="Invalid Category: %s" % category)
 
         # Organize by Categories / Commands
-        if category == "Systems":
+        if category == "Service":
+            # service-level commands are always available
+            for command in command_list:
+                if command == "CheckAvailability":
+                    result["service"] = rf_utils.check_service_availability()
+
+        elif category == "Systems":
             # execute only if we find a Systems resource
             resource = rf_utils._find_systems_resource()
             if resource['ret'] is False: