fix 1201: for a KDU with configuration, wait for KDU up instead of deployed at RO
[osm/LCM.git] / osm_lcm / lcm.py
index 8cd15e8..eaf2558 100644 (file)
@@ -29,12 +29,11 @@ import logging.handlers
 import getopt
 import sys
 
-from osm_lcm import ns
-from osm_lcm import vim_sdn
-from osm_lcm import netslice
-from osm_lcm import ROclient
+from osm_lcm import ns, prometheus, vim_sdn, netslice
+from osm_lcm.ng_ro import NgRoException, NgRoClient
+from osm_lcm.ROclient import ROClient, ROClientException
 
-from time import time, sleep
+from time import time
 from osm_lcm.lcm_utils import versiontuple, LcmException, TaskRegistry, LcmExceptionExit
 from osm_lcm import version as lcm_version, version_date as lcm_version_date
 
@@ -63,6 +62,8 @@ class Lcm:
 
     ping_interval_pace = 120  # how many time ping is send once is confirmed all is running
     ping_interval_boot = 5    # how many time ping is sent when booting
+    cfg_logger_name = {"message": "lcm.msg", "database": "lcm.db", "storage": "lcm.fs", "tsdb": "lcm.prometheus"}
+    # ^ contains for each section at lcm.cfg the used logger name
 
     def __init__(self, config_file, loop=None):
         """
@@ -86,11 +87,18 @@ class Lcm:
         config = self.read_config_file(config_file)
         self.config = config
         self.config["ro_config"] = {
-            "endpoint_url": "http://{}:{}/openmano".format(config["RO"]["host"], config["RO"]["port"]),
+            "ng": config["RO"].get("ng", False),
+            "uri": config["RO"].get("uri"),
             "tenant": config.get("tenant", "osm"),
-            "logger_name": "lcm.ROclient",
-            "loglevel": "ERROR",
+            "logger_name": "lcm.roclient",
+            "loglevel": config["RO"].get("loglevel", "ERROR"),
         }
+        if not self.config["ro_config"]["uri"]:
+            if not self.config["ro_config"]["ng"]:
+                self.config["ro_config"]["uri"] = "http://{}:{}/openmano".format(config["RO"]["host"],
+                                                                                 config["RO"]["port"])
+            else:
+                self.config["ro_config"]["uri"] = "http://{}:{}/ro".format(config["RO"]["host"], config["RO"]["port"])
 
         self.loop = loop or asyncio.get_event_loop()
 
@@ -114,7 +122,7 @@ class Lcm:
             self.logger.setLevel(config["global"]["loglevel"])
 
         # logging other modules
-        for k1, logname in {"message": "lcm.msg", "database": "lcm.db", "storage": "lcm.fs"}.items():
+        for k1, logname in self.cfg_logger_name.items():
             config[k1]["logger_name"] = logname
             logger_module = logging.getLogger(logname)
             if config[k1].get("logfile"):
@@ -184,8 +192,17 @@ class Lcm:
         # contains created tasks/futures to be able to cancel
         self.lcm_tasks = TaskRegistry(self.worker_id, self.db, self.logger)
 
-        self.ns = ns.NsLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop)
-        self.netslice = netslice.NetsliceLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop)
+        if self.config.get("tsdb") and self.config["tsdb"].get("driver"):
+            if self.config["tsdb"]["driver"] == "prometheus":
+                self.prometheus = prometheus.Prometheus(self.config["tsdb"], self.worker_id, self.db, self.loop)
+            else:
+                raise LcmException("Invalid configuration param '{}' at '[tsdb]':'driver'".format(
+                    config["tsdb"]["driver"]))
+        else:
+            self.prometheus = None
+        self.ns = ns.NsLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop, self.prometheus)
+        self.netslice = netslice.NetsliceLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop,
+                                             self.ns)
         self.vim = vim_sdn.VimLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop)
         self.wim = vim_sdn.WimLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop)
         self.sdn = vim_sdn.SdnLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.config, self.loop)
@@ -197,17 +214,19 @@ class Lcm:
         last_error = None
         while True:
             try:
-                ro_server = ROclient.ROClient(self.loop, **self.config["ro_config"])
+                if self.config["ro_config"].get("ng"):
+                    ro_server = NgRoClient(self.loop, **self.config["ro_config"])
+                else:
+                    ro_server = ROClient(self.loop, **self.config["ro_config"])
                 ro_version = await ro_server.get_version()
                 if versiontuple(ro_version) < versiontuple(min_RO_version):
                     raise LcmException("Not compatible osm/RO version '{}'. Needed '{}' or higher".format(
                         ro_version, min_RO_version))
                 self.logger.info("Connected to RO version {}".format(ro_version))
                 return
-            except ROclient.ROClientException as e:
+            except (ROClientException, NgRoException) as e:
                 tries -= 1
-                error_text = "Error while connecting to RO on {}: {}".format(self.config["ro_config"]["endpoint_url"],
-                                                                             e)
+                error_text = "Error while connecting to RO on {}: {}".format(self.config["ro_config"]["uri"], e)
                 if tries <= 0:
                     self.logger.critical(error_text)
                     raise LcmException(error_text)
@@ -286,6 +305,10 @@ class Lcm:
                 except Exception as e:
                     self.logger.error("Cannot write into '{}' for healthcheck: {}".format(health_check_file, e))
             return
+        elif topic == "pla":
+            if command == "placement":
+                self.ns.update_nsrs_with_pla_result(params)
+            return
         elif topic == "k8scluster":
             if command == "create" or command == "created":
                 k8scluster_id = params.get("_id")
@@ -463,10 +486,10 @@ class Lcm:
         self.first_start = True
         while self.consecutive_errors < 10:
             try:
-                topics = ("ns", "vim_account", "wim_account", "sdn", "nsi", "k8scluster", "k8srepo")
+                topics = ("ns", "vim_account", "wim_account", "sdn", "nsi", "k8scluster", "k8srepo", "pla")
                 topics_admin = ("admin", )
                 await asyncio.gather(
-                    self.msg.aioread(topics, self.loop, self.kafka_read_callback),
+                    self.msg.aioread(topics, self.loop, self.kafka_read_callback, from_beginning=True),
                     self.msg_admin.aioread(topics_admin, self.loop, self.kafka_read_callback, group_id=False)
                 )
 
@@ -492,6 +515,10 @@ class Lcm:
         # check RO version
         self.loop.run_until_complete(self.check_RO_version())
 
+        # configure tsdb prometheus
+        if self.prometheus:
+            self.loop.run_until_complete(self.prometheus.start())
+
         self.loop.run_until_complete(asyncio.gather(
             self.kafka_read(),
             self.kafka_ping()
@@ -604,27 +631,10 @@ def usage():
     # --log-socket-port PORT: send logs using this port (default: 9022)")
 
 
-def health_check():
-    retry = 2
-    while retry:
-        retry -= 1
-        try:
-            with open(health_check_file, "r") as f:
-                last_received_ping = f.read()
-
-            if time() - float(last_received_ping) < Lcm.ping_interval_pace + 10:
-                exit(0)
-        except Exception:
-            pass
-        if retry:
-            sleep(6)
-    exit(1)
-
-
 if __name__ == '__main__':
 
     try:
-        print("SYS.PATH='{}'".format(sys.path))
+        print("SYS.PATH='{}'".format(sys.path))
         # load parameters and configuration
         # -h
         # -c value
@@ -641,7 +651,8 @@ if __name__ == '__main__':
             elif o in ("-c", "--config"):
                 config_file = a
             elif o == "--health-check":
-                health_check()
+                from osm_lcm.lcm_hc import health_check
+                health_check(health_check_file, Lcm.ping_interval_pace)
             # elif o == "--log-socket-port":
             #     log_socket_port = a
             # elif o == "--log-socket-host":