osm_lcm/lcm.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-

##
# Copyright 2018 Telefonica S.A.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
#         http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
##

import asyncio
import yaml
import logging
import logging.handlers
import getopt
import sys

from osm_lcm import ROclient, ns, vim_sdn, netslice
from time import time, sleep
from osm_lcm.lcm_utils import versiontuple, LcmException, TaskRegistry, LcmExceptionExit
from osm_lcm import version as lcm_version, version_date as lcm_version_date

from osm_common import dbmemory, dbmongo, fslocal, msglocal, msgkafka
from osm_common import version as common_version
from osm_common.dbbase import DbException
from osm_common.fsbase import FsException
from osm_common.msgbase import MsgException
from os import environ, path
from random import choice as random_choice
from n2vc import version as n2vc_version


__author__ = "Alfonso Tierno"
min_RO_version = "6.0.2"
min_n2vc_version = "0.0.2"
min_common_version = "0.1.19"
# uncomment if LCM is installed as library and installed, and get them from __init__.py
# lcm_version = '0.1.41'
# lcm_version_date = '2019-06-19'
health_check_file = path.expanduser("~") + "/time_last_ping"   # TODO find better location for this file


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

    def __init__(self, config_file, loop=None):
        """
        Init, Connect to database, filesystem storage, and messaging
        :param config: two level dictionary with configuration. Top level should contain 'database', 'storage',
        :return: None
        """

        self.db = None
        self.msg = None
        self.msg_admin = None
        self.fs = None
        self.pings_not_received = 1
        self.consecutive_errors = 0
        self.first_start = False

        # logging
        self.logger = logging.getLogger('lcm')
        # get id
        self.worker_id = self.get_process_id()
        # load configuration
        config = self.read_config_file(config_file)
        self.config = config
        self.ro_config = {
            "endpoint_url": "http://{}:{}/openmano".format(config["RO"]["host"], config["RO"]["port"]),
            "tenant": config.get("tenant", "osm"),
            "logger_name": "lcm.ROclient",
            "loglevel": "ERROR",
        }

        self.vca_config = config["VCA"]

        self.loop = loop or asyncio.get_event_loop()

        # logging
        log_format_simple = "%(asctime)s %(levelname)s %(name)s %(filename)s:%(lineno)s %(message)s"
        log_formatter_simple = logging.Formatter(log_format_simple, datefmt='%Y-%m-%dT%H:%M:%S')
        config["database"]["logger_name"] = "lcm.db"
        config["storage"]["logger_name"] = "lcm.fs"
        config["message"]["logger_name"] = "lcm.msg"
        if config["global"].get("logfile"):
            file_handler = logging.handlers.RotatingFileHandler(config["global"]["logfile"],
                                                                maxBytes=100e6, backupCount=9, delay=0)
            file_handler.setFormatter(log_formatter_simple)
            self.logger.addHandler(file_handler)
        if not config["global"].get("nologging"):
            str_handler = logging.StreamHandler()
            str_handler.setFormatter(log_formatter_simple)
            self.logger.addHandler(str_handler)

        if config["global"].get("loglevel"):
            self.logger.setLevel(config["global"]["loglevel"])

        # logging other modules
        for k1, logname in {"message": "lcm.msg", "database": "lcm.db", "storage": "lcm.fs"}.items():
            config[k1]["logger_name"] = logname
            logger_module = logging.getLogger(logname)
            if config[k1].get("logfile"):
                file_handler = logging.handlers.RotatingFileHandler(config[k1]["logfile"],
                                                                    maxBytes=100e6, backupCount=9, delay=0)
                file_handler.setFormatter(log_formatter_simple)
                logger_module.addHandler(file_handler)
            if config[k1].get("loglevel"):
                logger_module.setLevel(config[k1]["loglevel"])
        self.logger.critical("starting osm/lcm version {} {}".format(lcm_version, lcm_version_date))

        # check version of N2VC
        # TODO enhance with int conversion or from distutils.version import LooseVersion
        # or with list(map(int, version.split(".")))
        if versiontuple(n2vc_version) < versiontuple(min_n2vc_version):
            raise LcmException("Not compatible osm/N2VC version '{}'. Needed '{}' or higher".format(
                n2vc_version, min_n2vc_version))
        # check version of common
        if versiontuple(common_version) < versiontuple(min_common_version):
            raise LcmException("Not compatible osm/common version '{}'. Needed '{}' or higher".format(
                common_version, min_common_version))

        try:
            # TODO check database version
            if config["database"]["driver"] == "mongo":
                self.db = dbmongo.DbMongo()
                self.db.db_connect(config["database"])
            elif config["database"]["driver"] == "memory":
                self.db = dbmemory.DbMemory()
                self.db.db_connect(config["database"])
            else:
                raise LcmException("Invalid configuration param '{}' at '[database]':'driver'".format(
                    config["database"]["driver"]))

            if config["storage"]["driver"] == "local":
                self.fs = fslocal.FsLocal()
                self.fs.fs_connect(config["storage"])
            else:
                raise LcmException("Invalid configuration param '{}' at '[storage]':'driver'".format(
                    config["storage"]["driver"]))

            config_message = config["message"].copy()
            config_message["loop"] = self.loop
            if config_message["driver"] == "local":
                self.msg = msglocal.MsgLocal()
                self.msg.connect(config_message)
                self.msg_admin = msglocal.MsgLocal()
                config_message.pop("group_id", None)
                self.msg_admin.connect(config_message)
            elif config_message["driver"] == "kafka":
                self.msg = msgkafka.MsgKafka()
                self.msg.connect(config_message)
                self.msg_admin = msgkafka.MsgKafka()
                config_message.pop("group_id", None)
                self.msg_admin.connect(config_message)
            else:
                raise LcmException("Invalid configuration param '{}' at '[message]':'driver'".format(
                    config["message"]["driver"]))
        except (DbException, FsException, MsgException) as e:
            self.logger.critical(str(e), exc_info=True)
            raise LcmException(str(e))

        # 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.ro_config, self.vca_config, self.loop)
        self.netslice = netslice.NetsliceLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.ro_config, 
                                             self.vca_config, self.loop)
        self.vim = vim_sdn.VimLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.ro_config, self.loop)
        self.wim = vim_sdn.WimLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.ro_config, self.loop)
        self.sdn = vim_sdn.SdnLcm(self.db, self.msg, self.fs, self.lcm_tasks, self.ro_config, self.loop)

    async def check_RO_version(self):
        tries = 14
        last_error = None
        while True:
            try:
                ro_server = ROclient.ROClient(self.loop, **self.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:
                tries -= 1
                error_text = "Error while connecting to RO on {}: {}".format(self.ro_config["endpoint_url"], e)
                if tries <= 0:
                    self.logger.critical(error_text)
                    raise LcmException(error_text)
                if last_error != error_text:
                    last_error = error_text
                    self.logger.error(error_text + ". Waiting until {} seconds".format(5*tries))
                await asyncio.sleep(5)

    async def test(self, param=None):
        self.logger.debug("Starting/Ending test task: {}".format(param))

    async def kafka_ping(self):
        self.logger.debug("Task kafka_ping Enter")
        consecutive_errors = 0
        first_start = True
        kafka_has_received = False
        self.pings_not_received = 1
        while True:
            try:
                await self.msg_admin.aiowrite(
                    "admin", "ping",
                    {"from": "lcm", "to": "lcm", "worker_id": self.worker_id, "version": lcm_version},
                    self.loop)
                # time between pings are low when it is not received and at starting
                wait_time = self.ping_interval_boot if not kafka_has_received else self.ping_interval_pace
                if not self.pings_not_received:
                    kafka_has_received = True
                self.pings_not_received += 1
                await asyncio.sleep(wait_time, loop=self.loop)
                if self.pings_not_received > 10:
                    raise LcmException("It is not receiving pings from Kafka bus")
                consecutive_errors = 0
                first_start = False
            except LcmException:
                raise
            except Exception as e:
                # if not first_start is the first time after starting. So leave more time and wait
                # to allow kafka starts
                if consecutive_errors == 8 if not first_start else 30:
                    self.logger.error("Task kafka_read task exit error too many errors. Exception: {}".format(e))
                    raise
                consecutive_errors += 1
                self.logger.error("Task kafka_read retrying after Exception {}".format(e))
                wait_time = 2 if not first_start else 5
                await asyncio.sleep(wait_time, loop=self.loop)

    def kafka_read_callback(self, topic, command, params):
        order_id = 1

        if topic != "admin" and command != "ping":
            self.logger.debug("Task kafka_read receives {} {}: {}".format(topic, command, params))
        self.consecutive_errors = 0
        self.first_start = False
        order_id += 1
        if command == "exit":
            raise LcmExceptionExit
        elif command.startswith("#"):
            return
        elif command == "echo":
            # just for test
            print(params)
            sys.stdout.flush()
            return
        elif command == "test":
            asyncio.Task(self.test(params), loop=self.loop)
            return

        if topic == "admin":
            if command == "ping" and params["to"] == "lcm" and params["from"] == "lcm":
                if params.get("worker_id") != self.worker_id:
                    return
                self.pings_not_received = 0
                try:
                    with open(health_check_file, "w") as f:
                        f.write(str(time()))
                except Exception as e:
                    self.logger.error("Cannot write into '{}' for healthcheck: {}".format(health_check_file, e))
            return
        elif topic == "ns":
            if command == "instantiate":
                # self.logger.debug("Deploying NS {}".format(nsr_id))
                nslcmop = params
                nslcmop_id = nslcmop["_id"]
                nsr_id = nslcmop["nsInstanceId"]
                task = asyncio.ensure_future(self.ns.instantiate(nsr_id, nslcmop_id))
                self.lcm_tasks.register("ns", nsr_id, nslcmop_id, "ns_instantiate", task)
                return
            elif command == "terminate":
                # self.logger.debug("Deleting NS {}".format(nsr_id))
                nslcmop = params
                nslcmop_id = nslcmop["_id"]
                nsr_id = nslcmop["nsInstanceId"]
                self.lcm_tasks.cancel(topic, nsr_id)
                task = asyncio.ensure_future(self.ns.terminate(nsr_id, nslcmop_id))
                self.lcm_tasks.register("ns", nsr_id, nslcmop_id, "ns_terminate", task)
                return
            elif command == "action":
                # self.logger.debug("Update NS {}".format(nsr_id))
                nslcmop = params
                nslcmop_id = nslcmop["_id"]
                nsr_id = nslcmop["nsInstanceId"]
                task = asyncio.ensure_future(self.ns.action(nsr_id, nslcmop_id))
                self.lcm_tasks.register("ns", nsr_id, nslcmop_id, "ns_action", task)
                return
            elif command == "scale":
                # self.logger.debug("Update NS {}".format(nsr_id))
                nslcmop = params
                nslcmop_id = nslcmop["_id"]
                nsr_id = nslcmop["nsInstanceId"]
                task = asyncio.ensure_future(self.ns.scale(nsr_id, nslcmop_id))
                self.lcm_tasks.register("ns", nsr_id, nslcmop_id, "ns_scale", task)
                return
            elif command == "show":
                nsr_id = params
                try:
                    db_nsr = self.db.get_one("nsrs", {"_id": nsr_id})
                    print("nsr:\n    _id={}\n    operational-status: {}\n    config-status: {}"
                          "\n    detailed-status: {}\n    deploy: {}\n    tasks: {}"
                          "".format(nsr_id, db_nsr["operational-status"], db_nsr["config-status"],
                                    db_nsr["detailed-status"],
                                    db_nsr["_admin"]["deployed"], self.lcm_ns_tasks.get(nsr_id)))
                except Exception as e:
                    print("nsr {} not found: {}".format(nsr_id, e))
                sys.stdout.flush()
                return
            elif command == "deleted":
                return  # TODO cleaning of task just in case should be done
            elif command in ("terminated", "instantiated", "scaled", "actioned"):  # "scaled-cooldown-time"
                return
        elif topic == "nsi":  # netslice LCM processes (instantiate, terminate, etc)
            if command == "instantiate":
                # self.logger.debug("Instantiating Network Slice {}".format(nsilcmop["netsliceInstanceId"]))
                nsilcmop = params
                nsilcmop_id = nsilcmop["_id"]  # slice operation id
                nsir_id = nsilcmop["netsliceInstanceId"]  # slice record id
                task = asyncio.ensure_future(self.netslice.instantiate(nsir_id, nsilcmop_id))
                self.lcm_tasks.register("nsi", nsir_id, nsilcmop_id, "nsi_instantiate", task)
                return
            elif command == "terminate":
                # self.logger.debug("Terminating Network Slice NS {}".format(nsilcmop["netsliceInstanceId"]))
                nsilcmop = params
                nsilcmop_id = nsilcmop["_id"]  # slice operation id
                nsir_id = nsilcmop["netsliceInstanceId"]  # slice record id
                self.lcm_tasks.cancel(topic, nsir_id)
                task = asyncio.ensure_future(self.netslice.terminate(nsir_id, nsilcmop_id))
                self.lcm_tasks.register("nsi", nsir_id, nsilcmop_id, "nsi_terminate", task)
                return
            elif command == "show":
                nsir_id = params
                try:
                    db_nsir = self.db.get_one("nsirs", {"_id": nsir_id})
                    print("nsir:\n    _id={}\n    operational-status: {}\n    config-status: {}"
                          "\n    detailed-status: {}\n    deploy: {}\n    tasks: {}"
                          "".format(nsir_id, db_nsir["operational-status"], db_nsir["config-status"],
                                    db_nsir["detailed-status"],
                                    db_nsir["_admin"]["deployed"], self.lcm_netslice_tasks.get(nsir_id)))
                except Exception as e:
                    print("nsir {} not found: {}".format(nsir_id, e))
                sys.stdout.flush()
                return
            elif command == "deleted":
                return  # TODO cleaning of task just in case should be done
            elif command in ("terminated", "instantiated", "scaled", "actioned"):  # "scaled-cooldown-time"
                return
        elif topic == "vim_account":
            vim_id = params["_id"]
            if command in ("create", "created"):
                task = asyncio.ensure_future(self.vim.create(params, order_id))
                self.lcm_tasks.register("vim_account", vim_id, order_id, "vim_create", task)
                return
            elif command == "delete":
                self.lcm_tasks.cancel(topic, vim_id)
                task = asyncio.ensure_future(self.vim.delete(params, order_id))
                self.lcm_tasks.register("vim_account", vim_id, order_id, "vim_delete", task)
                return
            elif command == "show":
                print("not implemented show with vim_account")
                sys.stdout.flush()
                return
            elif command in ("edit", "edited"):
                task = asyncio.ensure_future(self.vim.edit(params, order_id))
                self.lcm_tasks.register("vim_account", vim_id, order_id, "vim_edit", task)
                return
            elif command == "deleted":
                return  # TODO cleaning of task just in case should be done
        elif topic == "wim_account":
            wim_id = params["_id"]
            if command in ("create", "created"):
                task = asyncio.ensure_future(self.wim.create(params, order_id))
                self.lcm_tasks.register("wim_account", wim_id, order_id, "wim_create", task)
                return
            elif command == "delete":
                self.lcm_tasks.cancel(topic, wim_id)
                task = asyncio.ensure_future(self.wim.delete(params, order_id))
                self.lcm_tasks.register("wim_account", wim_id, order_id, "wim_delete", task)
                return
            elif command == "show":
                print("not implemented show with wim_account")
                sys.stdout.flush()
                return
            elif command in ("edit", "edited"):
                task = asyncio.ensure_future(self.wim.edit(params, order_id))
                self.lcm_tasks.register("wim_account", wim_id, order_id, "wim_edit", task)
                return
            elif command == "deleted":
                return  # TODO cleaning of task just in case should be done
        elif topic == "sdn":
            _sdn_id = params["_id"]
            if command in ("create", "created"):
                task = asyncio.ensure_future(self.sdn.create(params, order_id))
                self.lcm_tasks.register("sdn", _sdn_id, order_id, "sdn_create", task)
                return
            elif command == "delete":
                self.lcm_tasks.cancel(topic, _sdn_id)
                task = asyncio.ensure_future(self.sdn.delete(params, order_id))
                self.lcm_tasks.register("sdn", _sdn_id, order_id, "sdn_delete", task)
                return
            elif command in ("edit", "edited"):
                task = asyncio.ensure_future(self.sdn.edit(params, order_id))
                self.lcm_tasks.register("sdn", _sdn_id, order_id, "sdn_edit", task)
                return
            elif command == "deleted":
                return  # TODO cleaning of task just in case should be done
        self.logger.critical("unknown topic {} and command '{}'".format(topic, command))

    async def kafka_read(self):
        self.logger.debug("Task kafka_read Enter with worker_id={}".format(self.worker_id))
        # future = asyncio.Future()
        self.consecutive_errors = 0
        self.first_start = True
        while self.consecutive_errors < 10:
            try:
                topics = ("ns", "vim_account", "wim_account", "sdn", "nsi")
                topics_admin = ("admin", )
                await asyncio.gather(
                    self.msg.aioread(topics, self.loop, self.kafka_read_callback),
                    self.msg_admin.aioread(topics_admin, self.loop, self.kafka_read_callback, group_id=False)
                )

            except LcmExceptionExit:
                self.logger.debug("Bye!")
                break
            except Exception as e:
                # if not first_start is the first time after starting. So leave more time and wait
                # to allow kafka starts
                if self.consecutive_errors == 8 if not self.first_start else 30:
                    self.logger.error("Task kafka_read task exit error too many errors. Exception: {}".format(e))
                    raise
                self.consecutive_errors += 1
                self.logger.error("Task kafka_read retrying after Exception {}".format(e))
                wait_time = 2 if not self.first_start else 5
                await asyncio.sleep(wait_time, loop=self.loop)

        # self.logger.debug("Task kafka_read terminating")
        self.logger.debug("Task kafka_read exit")

    def start(self):

        # check RO version
        self.loop.run_until_complete(self.check_RO_version())

        self.loop.run_until_complete(asyncio.gather(
            self.kafka_read(),
            self.kafka_ping()
        ))
        # TODO
        # self.logger.debug("Terminating cancelling creation tasks")
        # self.lcm_tasks.cancel("ALL", "create")
        # timeout = 200
        # while self.is_pending_tasks():
        #     self.logger.debug("Task kafka_read terminating. Waiting for tasks termination")
        #     await asyncio.sleep(2, loop=self.loop)
        #     timeout -= 2
        #     if not timeout:
        #         self.lcm_tasks.cancel("ALL", "ALL")
        self.loop.close()
        self.loop = None
        if self.db:
            self.db.db_disconnect()
        if self.msg:
            self.msg.disconnect()
        if self.msg_admin:
            self.msg_admin.disconnect()
        if self.fs:
            self.fs.fs_disconnect()

    def read_config_file(self, config_file):
        # TODO make a [ini] + yaml inside parser
        # the configparser library is not suitable, because it does not admit comments at the end of line,
        # and not parse integer or boolean
        try:
            with open(config_file) as f:
                conf = yaml.load(f)
            for k, v in environ.items():
                if not k.startswith("OSMLCM_"):
                    continue
                k_items = k.lower().split("_")
                if len(k_items) < 3:
                    continue
                if k_items[1] in ("ro", "vca"):
                    # put in capital letter
                    k_items[1] = k_items[1].upper()
                c = conf
                try:
                    for k_item in k_items[1:-1]:
                        c = c[k_item]
                    if k_items[-1] == "port":
                        c[k_items[-1]] = int(v)
                    else:
                        c[k_items[-1]] = v
                except Exception as e:
                    self.logger.warn("skipping environ '{}' on exception '{}'".format(k, e))

            return conf
        except Exception as e:
            self.logger.critical("At config file '{}': {}".format(config_file, e))
            exit(1)

    @staticmethod
    def get_process_id():
        """
        Obtain a unique ID for this process. If running from inside docker, it will get docker ID. If not it
        will provide a random one
        :return: Obtained ID
        """
        # Try getting docker id. If fails, get pid
        try:
            with open("/proc/self/cgroup", "r") as f:
                text_id_ = f.readline()
                _, _, text_id = text_id_.rpartition("/")
                text_id = text_id.replace('\n', '')[:12]
                if text_id:
                    return text_id
        except Exception:
            pass
        # Return a random id
        return ''.join(random_choice("0123456789abcdef") for _ in range(12))


def usage():
    print("""Usage: {} [options]
        -c|--config [configuration_file]: loads the configuration file (default: ./nbi.cfg)
        --health-check: do not run lcm, but inspect kafka bus to determine if lcm is healthy
        -h|--help: shows this help
        """.format(sys.argv[0]))
    # --log-socket-host HOST: send logs to this host")
    # --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:
        # load parameters and configuration
        opts, args = getopt.getopt(sys.argv[1:], "hc:", ["config=", "help", "health-check"])
        # TODO add  "log-socket-host=", "log-socket-port=", "log-file="
        config_file = None
        for o, a in opts:
            if o in ("-h", "--help"):
                usage()
                sys.exit()
            elif o in ("-c", "--config"):
                config_file = a
            elif o == "--health-check":
                health_check()
            # elif o == "--log-socket-port":
            #     log_socket_port = a
            # elif o == "--log-socket-host":
            #     log_socket_host = a
            # elif o == "--log-file":
            #     log_file = a
            else:
                assert False, "Unhandled option"
        if config_file:
            if not path.isfile(config_file):
                print("configuration file '{}' not exist".format(config_file), file=sys.stderr)
                exit(1)
        else:
            for config_file in (__file__[:__file__.rfind(".")] + ".cfg", "./lcm.cfg", "/etc/osm/lcm.cfg"):
                if path.isfile(config_file):
                    break
            else:
                print("No configuration file 'lcm.cfg' found neither at local folder nor at /etc/osm/", file=sys.stderr)
                exit(1)
        lcm = Lcm(config_file)
        lcm.start()
    except (LcmException, getopt.GetoptError) as e:
        print(str(e), file=sys.stderr)
        # usage()
        exit(1)