Added new helm grpc connector 37/9237/10
authorlloretgalleg <illoret@indra.es>
Tue, 30 Jun 2020 13:46:28 +0000 (13:46 +0000)
committerlloretgalleg <illoret@indra.es>
Wed, 1 Jul 2020 14:14:10 +0000 (14:14 +0000)
Change-Id: I15b4a8fe28c679130017fc4982066f9a2de243d6
Signed-off-by: lloretgalleg <illoret@indra.es>
osm_lcm/lcm_helm_conn.py [new file with mode: 0644]
stdeb.cfg
tox.ini

diff --git a/osm_lcm/lcm_helm_conn.py b/osm_lcm/lcm_helm_conn.py
new file mode 100644 (file)
index 0000000..c7ea476
--- /dev/null
@@ -0,0 +1,484 @@
+##
+# Copyright 2020 Telefonica Investigacion y Desarrollo, S.A.U.
+#
+# 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 functools
+import yaml
+import asyncio
+import socket
+import uuid
+
+from grpclib.client import Channel
+
+from osm_lcm.frontend_pb2 import PrimitiveRequest
+from osm_lcm.frontend_pb2 import SshKeyRequest, SshKeyReply
+from osm_lcm.frontend_grpc import FrontendExecutorStub
+
+from n2vc.n2vc_conn import N2VCConnector
+from n2vc.k8s_helm_conn import K8sHelmConnector
+from n2vc.exceptions import N2VCBadArgumentsException, N2VCException, N2VCExecutionException
+
+from osm_lcm.lcm_utils import deep_get
+
+
+def retryer(max_wait_time=60, delay_time=10):
+    def wrapper(func):
+        retry_exceptions = (
+            ConnectionRefusedError
+        )
+
+        @functools.wraps(func)
+        async def wrapped(*args, **kwargs):
+            wait_time = max_wait_time
+            while wait_time > 0:
+                try:
+                    return await func(*args, **kwargs)
+                except retry_exceptions:
+                    wait_time = wait_time - delay_time
+                    await asyncio.sleep(delay_time)
+                    continue
+            else:
+                return ConnectionRefusedError
+        return wrapped
+    return wrapper
+
+
+class LCMHelmConn(N2VCConnector):
+    _KUBECTL_OSM_NAMESPACE = "osm"
+    _KUBECTL_OSM_CLUSTER_NAME = "_system-osm-k8s"
+    _EE_SERVICE_PORT = 50050
+
+    # Time beetween retries
+    _EE_RETRY_DELAY = 10
+    # Initial max retry time
+    _MAX_INITIAL_RETRY_TIME = 300
+    # Other retry time
+    _MAX_RETRY_TIME = 30
+
+    def __init__(self,
+                 db: object,
+                 fs: object,
+                 log: object = None,
+                 loop: object = None,
+                 url: str = None,
+                 username: str = None,
+                 vca_config: dict = None,
+                 on_update_db=None, ):
+        """
+        Initialize EE helm connector.
+        """
+
+        # parent class constructor
+        N2VCConnector.__init__(
+            self,
+            db=db,
+            fs=fs,
+            log=log,
+            loop=loop,
+            url=url,
+            username=username,
+            vca_config=vca_config,
+            on_update_db=on_update_db,
+        )
+
+        self.log.debug("Initialize helm N2VC connector")
+
+        # TODO - Obtain data from configuration
+        self._ee_service_port = self._EE_SERVICE_PORT
+
+        self._retry_delay = self._EE_RETRY_DELAY
+        self._max_retry_time = self._MAX_RETRY_TIME
+        self._initial_retry_time = self._MAX_INITIAL_RETRY_TIME
+
+        # initialize helm connector
+        self._k8sclusterhelm = K8sHelmConnector(
+            kubectl_command=self.vca_config.get("kubectlpath"),
+            helm_command=self.vca_config.get("helmpath"),
+            fs=self.fs,
+            log=self.log,
+            db=self.db,
+            on_update_db=None,
+        )
+
+        self._system_cluster_id = None
+        self.log.info("Helm N2VC connector initialized")
+
+    # TODO - ¿reuse_ee_id?
+    async def create_execution_environment(self,
+                                           namespace: str,
+                                           db_dict: dict,
+                                           reuse_ee_id: str = None,
+                                           progress_timeout: float = None,
+                                           total_timeout: float = None,
+                                           artifact_path: str = None,
+                                           vca_type: str = None) -> (str, dict):
+        """
+        Creates a new helm execution environment deploying the helm-chat indicated in the
+        attifact_path
+        :param str namespace: This param is not used, all helm charts are deployed in the osm
+        system namespace
+        :param dict db_dict: where to write to database when the status changes.
+            It contains a dictionary with {collection: str, filter: {},  path: str},
+                e.g. {collection: "nsrs", filter: {_id: <nsd-id>, path:
+                "_admin.deployed.VCA.3"}
+        :param str reuse_ee_id: ee id from an older execution. TODO - right now this params is not used
+        :param float progress_timeout:
+        :param float total_timeout:
+        :param str artifact_path  path of package content
+        :param str vca_type  Type of vca, not used as assumed of type helm
+        :returns str, dict: id of the new execution environment including namespace.helm_id
+        and credentials object set to None as all credentials should be osm kubernetes .kubeconfig
+        """
+
+        self.log.info(
+            "create_execution_environment: namespace: {}, artifact_path: {}, db_dict: {}, "
+            "reuse_ee_id: {}".format(
+                namespace, artifact_path, db_dict, reuse_ee_id)
+        )
+
+        # Validate artifact-path is provided
+        if artifact_path is None or len(artifact_path) == 0:
+            raise N2VCBadArgumentsException(
+                message="artifact_path is mandatory", bad_args=["artifact_path"]
+            )
+
+        # Validate artifact-path exists
+
+        # remove / in charm path
+        while artifact_path.find("//") >= 0:
+            artifact_path = artifact_path.replace("//", "/")
+
+        # check charm path
+        if self.fs.file_exists(artifact_path):
+            helm_chart_path = artifact_path
+        else:
+            msg = "artifact path does not exist: {}".format(artifact_path)
+            raise N2VCBadArgumentsException(message=msg, bad_args=["artifact_path"])
+
+        if artifact_path.startswith("/"):
+            full_path = self.fs.path + helm_chart_path
+        else:
+            full_path = self.fs.path + "/" + helm_chart_path
+
+        try:
+            # Call helm conn install
+            # Obtain system cluster id from database
+            system_cluster_uuid = self._get_system_cluster_id()
+
+            self.log.debug("install helm chart: {}".format(full_path))
+            helm_id = await self._k8sclusterhelm.install(system_cluster_uuid, kdu_model=full_path,
+                                                         namespace=self._KUBECTL_OSM_NAMESPACE,
+                                                         db_dict=db_dict,
+                                                         timeout=progress_timeout)
+
+            ee_id = "{}.{}".format(self._KUBECTL_OSM_NAMESPACE, helm_id)
+            return ee_id, None
+        except Exception as e:
+            self.log.error("Error deploying chart ee: {}".format(e), exc_info=True)
+            raise N2VCException("Error deploying chart ee: {}".format(e))
+
+    async def register_execution_environment(self, namespace: str, credentials: dict, db_dict: dict,
+                                             progress_timeout: float = None, total_timeout: float = None) -> str:
+        # nothing to do
+        pass
+
+    async def install_configuration_sw(self,
+                                       ee_id: str,
+                                       artifact_path: str,
+                                       db_dict: dict,
+                                       progress_timeout: float = None,
+                                       total_timeout: float = None,
+                                       config: dict = None,
+                                       ):
+        # nothing to do
+        pass
+
+    async def add_relation(self, ee_id_1: str, ee_id_2: str, endpoint_1: str, endpoint_2: str):
+        # nothing to do
+        pass
+
+    async def remove_relation(self):
+        # nothing to to
+        pass
+
+    async def get_status(self, namespace: str, yaml_format: bool = True):
+        # not used for this connector
+        pass
+
+    async def get_ee_ssh_public__key(self, ee_id: str, db_dict: dict, progress_timeout: float = None,
+                                     total_timeout: float = None) -> str:
+        """
+        Obtains ssh-public key from ee executing GetSShKey method from the ee.
+
+        :param str ee_id: the id of the execution environment returned by
+            create_execution_environment or register_execution_environment
+        :param dict db_dict:
+        :param float progress_timeout:
+        :param float total_timeout:
+        :returns: public key of the execution environment
+        """
+
+        self.log.info(
+            "get_ee_ssh_public_key: ee_id: {}, db_dict: {}".format(
+                ee_id, db_dict)
+        )
+
+        # check arguments
+        if ee_id is None or len(ee_id) == 0:
+            raise N2VCBadArgumentsException(
+                message="ee_id is mandatory", bad_args=["ee_id"]
+            )
+
+        try:
+            # Obtain ip_addr for the ee service, it is resolved by dns from the ee name by kubernetes
+            namespace, helm_id = self._get_ee_id_parts(ee_id)
+            ip_addr = socket.gethostbyname(helm_id)
+
+            # Obtain ssh_key from the ee, this method will implement retries to allow the ee
+            # install libraries and start successfully
+            ssh_key = await self._get_ssh_key(ip_addr)
+            return ssh_key
+        except Exception as e:
+            self.log.error("Error obtaining ee ssh_key: {}".format(e), exc_info=True)
+            raise N2VCException("Error obtaining ee ssh_ke: {}".format(e))
+
+    async def exec_primitive(self, ee_id: str, primitive_name: str, params_dict: dict, db_dict: dict = None,
+                             progress_timeout: float = None, total_timeout: float = None) -> str:
+        """
+        Execute a primitive in the execution environment
+
+        :param str ee_id: the one returned by create_execution_environment or
+            register_execution_environment with the format namespace.helm_id
+        :param str primitive_name: must be one defined in the software. There is one
+            called 'config', where, for the proxy case, the 'credentials' of VM are
+            provided
+        :param dict params_dict: parameters of the action
+        :param dict db_dict: where to write into database when the status changes.
+                        It contains a dict with
+                            {collection: <str>, filter: {},  path: <str>},
+                            e.g. {collection: "nslcmops", filter:
+                                {_id: <nslcmop_id>, path: "_admin.VCA"}
+                        It will be used to store information about intermediate notifications
+        :param float progress_timeout:
+        :param float total_timeout:
+        :returns str: primitive result, if ok. It raises exceptions in case of fail
+        """
+
+        self.log.info("exec primitive for ee_id : {}, primitive_name: {}, params_dict: {}, db_dict: {}".format(
+            ee_id, primitive_name, params_dict, db_dict
+        ))
+
+        # check arguments
+        if ee_id is None or len(ee_id) == 0:
+            raise N2VCBadArgumentsException(
+                message="ee_id is mandatory", bad_args=["ee_id"]
+            )
+        if primitive_name is None or len(primitive_name) == 0:
+            raise N2VCBadArgumentsException(
+                message="action_name is mandatory", bad_args=["action_name"]
+            )
+        if params_dict is None:
+            params_dict = dict()
+
+        try:
+            namespace, helm_id = self._get_ee_id_parts(ee_id)
+            ip_addr = socket.gethostbyname(helm_id)
+        except Exception as e:
+            self.log.error("Error getting ee ip ee: {}".format(e))
+            raise N2VCException("Error getting ee ip ee: {}".format(e))
+
+        if primitive_name == "config":
+            try:
+                # Execute config primitive, higher timeout to check the case ee is starting
+                status, detailed_message = await self._execute_config_primitive(ip_addr, params_dict, db_dict=db_dict)
+                self.log.debug("Executed config primitive ee_id_ {}, status: {}, message: {}".format(
+                    ee_id, status, detailed_message))
+                if status != "OK":
+                    self.log.error("Error configuring helm ee, status: {}, message: {}".format(
+                        status, detailed_message))
+                    raise N2VCExecutionException(
+                        message="Error configuring helm ee_id: {}, status: {}, message: {}: ".format(
+                            ee_id, status, detailed_message
+                        ),
+                        primitive_name=primitive_name,
+                    )
+            except Exception as e:
+                self.log.error("Error configuring helm ee: {}".format(e))
+                raise N2VCExecutionException(
+                    message="Error configuring helm ee_id: {}, {}".format(
+                        ee_id, e
+                    ),
+                    primitive_name=primitive_name,
+                )
+            return "CONFIG OK"
+        else:
+            try:
+                # Execute primitive
+                status, detailed_message = await self._execute_primitive(ip_addr, primitive_name,
+                                                                         params_dict, db_dict=db_dict)
+                self.log.debug("Executed primitive {} ee_id_ {}, status: {}, message: {}".format(
+                    primitive_name, ee_id, status, detailed_message))
+                if status != "OK" and status != "PROCESSING":
+                    self.log.error(
+                        "Execute primitive {} returned not ok status: {}, message: {}".format(
+                            primitive_name, status, detailed_message)
+                    )
+                    raise N2VCExecutionException(
+                        message="Execute primitive {} returned not ok status: {}, message: {}".format(
+                            primitive_name, status, detailed_message
+                        ),
+                        primitive_name=primitive_name,
+                    )
+            except Exception as e:
+                self.log.error(
+                    "Error executing primitive {}: {}".format(primitive_name, e)
+                )
+                raise N2VCExecutionException(
+                    message="Error executing primitive {} into ee={} : {}".format(
+                        primitive_name, ee_id, e
+                    ),
+                    primitive_name=primitive_name,
+                )
+            return detailed_message
+
+    async def deregister_execution_environments(self):
+        # nothing to be done
+        pass
+
+    async def delete_execution_environment(self, ee_id: str, db_dict: dict = None, total_timeout: float = None):
+        """
+        Delete an execution environment
+        :param str ee_id: id of the execution environment to delete, included namespace.helm_id
+        :param dict db_dict: where to write into database when the status changes.
+                        It contains a dict with
+                            {collection: <str>, filter: {},  path: <str>},
+                            e.g. {collection: "nsrs", filter:
+                                {_id: <nsd-id>, path: "_admin.deployed.VCA.3"}
+        :param float total_timeout:
+        """
+
+        self.log.info("ee_id: {}".format(ee_id))
+
+        # check arguments
+        if ee_id is None:
+            raise N2VCBadArgumentsException(
+                message="ee_id is mandatory", bad_args=["ee_id"]
+            )
+
+        try:
+
+            # Obtain cluster_uuid
+            system_cluster_uuid = self._get_system_cluster_id()
+
+            # Get helm_id
+            namespace, helm_id = self._get_ee_id_parts(ee_id)
+
+            # Uninstall chart
+            await self._k8sclusterhelm.uninstall(system_cluster_uuid, helm_id)
+            self.log.info("ee_id: {} deleted".format(ee_id))
+        except Exception as e:
+            self.log.error("Error deleting ee id: {}: {}".format(ee_id, e), exc_info=True)
+            raise N2VCException("Error deleting ee id {}: {}".format(ee_id, e))
+
+    async def delete_namespace(self, namespace: str, db_dict: dict = None, total_timeout: float = None):
+        # method not implemented for this connector, execution environments must be deleted individually
+        pass
+
+    async def install_k8s_proxy_charm(
+        self,
+        charm_name: str,
+        namespace: str,
+        artifact_path: str,
+        db_dict: dict,
+        progress_timeout: float = None,
+        total_timeout: float = None,
+        config: dict = None,
+    ) -> str:
+        pass
+
+    @retryer(max_wait_time=_MAX_INITIAL_RETRY_TIME, delay_time=_EE_RETRY_DELAY)
+    async def _get_ssh_key(self, ip_addr):
+        channel = Channel(ip_addr, self._ee_service_port)
+        try:
+            stub = FrontendExecutorStub(channel)
+            self.log.debug("get ssh key, ip_addr: {}".format(ip_addr))
+            reply: SshKeyReply = await stub.GetSshKey(SshKeyRequest())
+            return reply.message
+        finally:
+            channel.close()
+
+    @retryer(max_wait_time=_MAX_INITIAL_RETRY_TIME, delay_time=_EE_RETRY_DELAY)
+    async def _execute_config_primitive(self, ip_addr, params, db_dict=None):
+        return await self._execute_primitive_internal(ip_addr, "config", params, db_dict=db_dict)
+
+    @retryer(max_wait_time=_MAX_RETRY_TIME, delay_time=_EE_RETRY_DELAY)
+    async def _execute_primitive(self, ip_addr, primitive_name, params, db_dict=None):
+        return await  self._execute_primitive_internal(ip_addr, primitive_name, params, db_dict=db_dict)
+
+    async def _execute_primitive_internal(self, ip_addr, primitive_name, params, db_dict=None):
+
+        channel = Channel(ip_addr, self._ee_service_port)
+        try:
+            stub = FrontendExecutorStub(channel)
+            async with stub.RunPrimitive.open() as stream:
+                primitive_id = str(uuid.uuid1())
+                result = None
+                self.log.debug("Execute primitive internal: id:{}, name:{}, params: {}".
+                               format(primitive_id, primitive_name, params))
+                await stream.send_message(
+                    PrimitiveRequest(id=primitive_id, name=primitive_name, params=yaml.dump(params)), end=True)
+                async for reply in stream:
+                    self.log.debug("Received reply: {}".format(reply))
+                    result = reply
+                    # If db_dict provided write notifs in database
+                    if db_dict:
+                        self._write_op_detailed_status(db_dict, reply.status, reply.detailed_message)
+                if result:
+                    return reply.status, reply.detailed_message
+                else:
+                    return "ERROR", "No result received"
+        finally:
+            channel.close()
+
+    def _write_op_detailed_status(self, db_dict, status, detailed_message):
+
+        # write ee_id to database: _admin.deployed.VCA.x
+        try:
+            the_table = db_dict["collection"]
+            the_filter = db_dict["filter"]
+            update_dict = {"detailed-status": "{}: {}".format(status, detailed_message)}
+            # self.log.debug('Writing ee_id to database: {}'.format(the_path))
+            self.db.set_one(
+                table=the_table,
+                q_filter=the_filter,
+                update_dict=update_dict,
+                fail_on_empty=True,
+            )
+        except asyncio.CancelledError:
+            raise
+        except Exception as e:
+            self.log.error("Error writing detailedStatus to database: {}".format(e))
+
+    def _get_system_cluster_id(self):
+        if not self._system_cluster_id:
+            db_k8cluster = self.db.get_one("k8sclusters", {"name": self._KUBECTL_OSM_CLUSTER_NAME})
+            k8s_hc_id = deep_get(db_k8cluster, ("_admin", "helm-chart", "id"))
+            self._system_cluster_id = k8s_hc_id
+        return self._system_cluster_id
+
+    def _get_ee_id_parts(self, ee_id):
+        namespace, _, helm_id = ee_id.partition('.')
+        return namespace, helm_id
index 3e222e5..2794a1f 100644 (file)
--- a/stdeb.cfg
+++ b/stdeb.cfg
@@ -16,5 +16,5 @@
 #
 ##
 [DEFAULT]
-X-Python3-Version : >= 3.5
+X-Python3-Version : >= 3.6
 Depends3: python3-osm-common, python3-n2vc, python3-yaml, python3-aiohttp, python3-jinja2, python3-pip
diff --git a/tox.ini b/tox.ini
index 24b770f..65a62e5 100644 (file)
--- a/tox.ini
+++ b/tox.ini
@@ -24,7 +24,7 @@ deps = -r{toxinidir}/test-requirements.txt
 commands=python3 -m unittest discover -v
 
 [testenv:flake8]
-basepython = python3
+basepython = python3.6
 deps = flake8
 commands = flake8 osm_lcm --max-line-length 120 \
     --exclude .svn,CVS,.gz,.git,__pycache__,.tox,local,temp,frontend_grpc.py,frontend_pb2.py \