Bug 2315: Fix for Unable to update healing policy using NS Update
[osm/POL.git] / osm_policy_module / core / agent.py
index 350f383..4a20527 100644 (file)
 ##
 import asyncio
 import logging
+from pathlib import Path
+import os
 
 import peewee
 
 from osm_policy_module.alarming.service import AlarmingService
 from osm_policy_module.autoscaling.service import AutoscalingService
+from osm_policy_module.healing.service import HealingService
+from osm_policy_module.common.common_db_client import CommonDbClient
 from osm_policy_module.common.message_bus_client import MessageBusClient
 from osm_policy_module.core.config import Config
 
 log = logging.getLogger(__name__)
 
-ALLOWED_KAFKA_KEYS = ['instantiated', 'scaled', 'terminated', 'notify_alarm']
+ALLOWED_KAFKA_KEYS = [
+    "instantiated",
+    "scaled",
+    "terminated",
+    "notify_alarm",
+    "policy_updated",
+    "vnf_terminated",
+]
 
 
 class PolicyModuleAgent:
-    def __init__(self, config: Config, loop=None):
+    def __init__(self, config: Config):
         self.conf = config
-        if not loop:
-            loop = asyncio.get_event_loop()
-        self.loop = loop
         self.msg_bus = MessageBusClient(config)
-        self.autoscaling_service = AutoscalingService(config, loop)
-        self.alarming_service = AlarmingService(config, loop)
+        self.db_client = CommonDbClient(config)
+        self.autoscaling_service = AutoscalingService(config)
+        self.alarming_service = AlarmingService(config)
+        self.healing_service = HealingService(config)
 
     def run(self):
-        self.loop.run_until_complete(self.start())
+        asyncio.run(self.start())
 
     async def start(self):
-        topics = [
-            "ns",
-            "alarm_response"
-        ]
+        Path("/tmp/osm_pol_agent_health_flag").touch()
+        topics = ["ns", "alarm_response"]
         await self.msg_bus.aioread(topics, self._process_msg)
         log.critical("Exiting...")
+        if os.path.exists("/tmp/osm_pol_agent_health_flag"):
+            os.remove("/tmp/osm_pol_agent_health_flag")
 
     async def _process_msg(self, topic, key, msg):
+        Path("/tmp/osm_pol_agent_health_flag").touch()
         log.debug("_process_msg topic=%s key=%s msg=%s", topic, key, msg)
         try:
             if key in ALLOWED_KAFKA_KEYS:
-
-                if key == 'instantiated':
+                if key == "instantiated":
                     await self._handle_instantiated(msg)
 
-                if key == 'scaled':
+                if key == "scaled":
                     await self._handle_scaled(msg)
 
-                if key == 'terminated':
+                if key == "terminated":
                     await self._handle_terminated(msg)
 
-                if key == 'notify_alarm':
+                if key == "notify_alarm":
                     await self._handle_alarm_notification(msg)
+
+                if key == "policy_updated":
+                    await self._handle_policy_update(msg)
+
+                if key == "vnf_terminated":
+                    await self._handle_vnf_terminated(msg)
             else:
                 log.debug("Key %s is not in ALLOWED_KAFKA_KEYS", key)
         except peewee.PeeweeException:
@@ -83,51 +99,124 @@ class PolicyModuleAgent:
 
     async def _handle_alarm_notification(self, content):
         log.debug("_handle_alarm_notification: %s", content)
-        alarm_uuid = content['notify_details']['alarm_uuid']
-        status = content['notify_details']['status']
+        alarm_uuid = content["notify_details"]["alarm_uuid"]
+        status = content["notify_details"]["status"]
         await self.autoscaling_service.handle_alarm(alarm_uuid, status)
         await self.alarming_service.handle_alarm(alarm_uuid, status, content)
+        await self.healing_service.handle_alarm(alarm_uuid, status)
 
     async def _handle_instantiated(self, content):
         log.debug("_handle_instantiated: %s", content)
-        nslcmop_id = content['nslcmop_id']
-        nslcmop = self.autoscaling_service.get_nslcmop(nslcmop_id)
-        if nslcmop['operationState'] == 'COMPLETED' or nslcmop['operationState'] == 'PARTIALLY_COMPLETED':
-            nsr_id = nslcmop['nsInstanceId']
+        nslcmop_id = content["nslcmop_id"]
+        nslcmop = self.db_client.get_nslcmop(nslcmop_id)
+        if (
+            nslcmop["operationState"] == "COMPLETED"
+            or nslcmop["operationState"] == "PARTIALLY_COMPLETED"
+        ):
+            nsr_id = nslcmop["nsInstanceId"]
             log.info("Configuring nsr_id: %s", nsr_id)
             await self.autoscaling_service.configure_scaling_groups(nsr_id)
             await self.alarming_service.configure_vnf_alarms(nsr_id)
+            await self.healing_service.configure_healing_alarms(nsr_id)
         else:
             log.info(
                 "Network_service is not in COMPLETED or PARTIALLY_COMPLETED state. "
                 "Current state is %s. Skipping...",
-                nslcmop['operationState'])
+                nslcmop["operationState"],
+            )
 
     async def _handle_scaled(self, content):
         log.debug("_handle_scaled: %s", content)
-        nslcmop_id = content['nslcmop_id']
-        nslcmop = self.autoscaling_service.get_nslcmop(nslcmop_id)
-        if nslcmop['operationState'] == 'COMPLETED' or nslcmop['operationState'] == 'PARTIALLY_COMPLETED':
-            nsr_id = nslcmop['nsInstanceId']
+        nslcmop_id = content["nslcmop_id"]
+        nslcmop = self.db_client.get_nslcmop(nslcmop_id)
+        if (
+            nslcmop["operationState"] == "COMPLETED"
+            or nslcmop["operationState"] == "PARTIALLY_COMPLETED"
+        ):
+            nsr_id = nslcmop["nsInstanceId"]
             log.info("Configuring scaled service with nsr_id: %s", nsr_id)
             await self.autoscaling_service.configure_scaling_groups(nsr_id)
             await self.autoscaling_service.delete_orphaned_alarms(nsr_id)
             await self.alarming_service.configure_vnf_alarms(nsr_id)
+            await self.healing_service.configure_healing_alarms(nsr_id)
+            await self.healing_service.delete_orphaned_healing_alarms(nsr_id)
         else:
             log.debug(
                 "Network service is not in COMPLETED or PARTIALLY_COMPLETED state. "
                 "Current state is %s. Skipping...",
-                nslcmop['operationState'])
+                nslcmop["operationState"],
+            )
 
     async def _handle_terminated(self, content):
         log.debug("_handle_deleted: %s", content)
-        nsr_id = content['nsr_id']
-        if content['operationState'] == 'COMPLETED' or content['operationState'] == 'PARTIALLY_COMPLETED':
-            log.info("Deleting scaling groups and alarms for network autoscaling_service with nsr_id: %s", nsr_id)
+        nsr_id = content["nsr_id"]
+        if (
+            content["operationState"] == "COMPLETED"
+            or content["operationState"] == "PARTIALLY_COMPLETED"
+        ):
+            log.info(
+                "Deleting scaling groups and alarms for network autoscaling_service with nsr_id: %s",
+                nsr_id,
+            )
             await self.autoscaling_service.delete_scaling_groups(nsr_id)
             await self.alarming_service.delete_vnf_alarms(nsr_id)
+            await self.healing_service.delete_healing_alarms(nsr_id)
+        else:
+            log.info(
+                "Network service is not in COMPLETED or PARTIALLY_COMPLETED state. "
+                "Current state is %s. Skipping...",
+                content["operationState"],
+            )
+
+    async def _handle_policy_update(self, content):
+        log.info("_handle_policy_update: %s", content)
+        nsr_id = content["nsr_id"]
+        vnf_member_index = content["vnf_member_index"]
+        if (
+            content["operationState"] == "COMPLETED"
+            or content["operationState"] == "PARTIALLY_COMPLETED"
+        ):
+            log.info(
+                "Updating policies of VNF with nsr_id: %s and vnf-member-index: %s"
+                % (nsr_id, vnf_member_index)
+            )
+            await self.autoscaling_service.delete_scaling_groups(
+                nsr_id, vnf_member_index
+            )
+            await self.alarming_service.delete_vnf_alarms(nsr_id, vnf_member_index)
+            await self.healing_service.delete_healing_alarms(nsr_id, vnf_member_index)
+            await self.autoscaling_service.configure_scaling_groups(
+                nsr_id, vnf_member_index
+            )
+            await self.alarming_service.configure_vnf_alarms(nsr_id, vnf_member_index)
+            await self.healing_service.configure_healing_alarms(
+                nsr_id, vnf_member_index
+            )
+        else:
+            log.info(
+                "Network service is not in COMPLETED or PARTIALLY_COMPLETED state. "
+                "Current state is %s. Skipping...",
+                content["operationState"],
+            )
+
+    async def _handle_vnf_terminated(self, content):
+        nsr_id = content["nsr_id"]
+        vnf_member_index = content["vnf_member_index"]
+        if (
+            content["operationState"] == "COMPLETED"
+            or content["operationState"] == "PARTIALLY_COMPLETED"
+        ):
+            log.info(
+                "Deleting policies of VNF with nsr_id: %s and vnf-member-index: %s"
+                % (nsr_id, vnf_member_index)
+            )
+            await self.autoscaling_service.delete_scaling_groups(
+                nsr_id, vnf_member_index
+            )
+            await self.alarming_service.delete_vnf_alarms(nsr_id, vnf_member_index)
         else:
             log.info(
                 "Network service is not in COMPLETED or PARTIALLY_COMPLETED state. "
                 "Current state is %s. Skipping...",
-                content['operationState'])
+                content["operationState"],
+            )