X-Git-Url: https://osm.etsi.org/gitweb/?a=blobdiff_plain;f=osm_lcm%2Fns.py;h=63ae2a409915d3c826b36cd7b8ee15ab74bf87ce;hb=9e9a923448e6673fc0396265732115b55f1df9da;hp=dd7ff3e7b30e98b072fb70043cb4538b6ce46c20;hpb=c773744f338d3c019b4978b0fc591d14c614b4cf;p=osm%2FLCM.git diff --git a/osm_lcm/ns.py b/osm_lcm/ns.py index dd7ff3e..63ae2a4 100644 --- a/osm_lcm/ns.py +++ b/osm_lcm/ns.py @@ -53,7 +53,6 @@ from osm_lcm.data_utils.vca import ( from osm_lcm.ng_ro import NgRoClient, NgRoException from osm_lcm.lcm_utils import ( LcmException, - LcmExceptionNoMgmtIP, LcmBase, deep_get, get_iterable, @@ -124,7 +123,7 @@ from copy import copy, deepcopy from time import time from uuid import uuid4 -from random import randint +from random import SystemRandom __author__ = "Alfonso Tierno " @@ -133,9 +132,18 @@ class NsLcm(LcmBase): SUBOPERATION_STATUS_NOT_FOUND = -1 SUBOPERATION_STATUS_NEW = -2 SUBOPERATION_STATUS_SKIP = -3 + EE_TLS_NAME = "ee-tls" task_name_deploy_vca = "Deploying VCA" - - def __init__(self, msg, lcm_tasks, config: LcmCfg, loop): + rel_operation_types = { + "GE": ">=", + "LE": "<=", + "GT": ">", + "LT": "<", + "EQ": "==", + "NE": "!=", + } + + def __init__(self, msg, lcm_tasks, config: LcmCfg): """ Init, Connect to database, filesystem storage, and messaging :param config: two level dictionary with configuration. Top level should contain 'database', 'storage', @@ -145,7 +153,6 @@ class NsLcm(LcmBase): self.db = Database().instance.db self.fs = Filesystem().instance.fs - self.loop = loop self.lcm_tasks = lcm_tasks self.timeout = config.timeout self.ro_config = config.RO @@ -154,7 +161,6 @@ class NsLcm(LcmBase): # create N2VC connector self.n2vc = N2VCJujuConnector( log=self.logger, - loop=self.loop, on_update_db=self._on_update_n2vc_db, fs=self.fs, db=self.db, @@ -162,7 +168,6 @@ class NsLcm(LcmBase): self.conn_helm_ee = LCMHelmConn( log=self.logger, - loop=self.loop, vca_config=self.vca_config, on_update_db=self._on_update_n2vc_db, ) @@ -189,7 +194,6 @@ class NsLcm(LcmBase): kubectl_command=self.vca_config.kubectlpath, juju_command=self.vca_config.jujupath, log=self.logger, - loop=self.loop, on_update_db=self._on_update_k8s_db, fs=self.fs, db=self.db, @@ -212,7 +216,7 @@ class NsLcm(LcmBase): } # create RO client - self.RO = NgRoClient(self.loop, **self.ro_config.to_dict()) + self.RO = NgRoClient(**self.ro_config.to_dict()) self.op_status_map = { "instantiation": self.RO.status, @@ -245,23 +249,6 @@ class NsLcm(LcmBase): pass return None - def _on_update_ro_db(self, nsrs_id, ro_descriptor): - # self.logger.debug('_on_update_ro_db(nsrs_id={}'.format(nsrs_id)) - - try: - # TODO filter RO descriptor fields... - - # write to database - db_dict = dict() - # db_dict['deploymentStatus'] = yaml.dump(ro_descriptor, default_flow_style=False, indent=2) - db_dict["deploymentStatus"] = ro_descriptor - self.update_db_2("nsrs", nsrs_id, db_dict) - - except Exception as e: - self.logger.warn( - "Cannot write database RO deployment for ns={} -> {}".format(nsrs_id, e) - ) - async def _on_update_n2vc_db(self, table, filter, path, updated_data, vca_id=None): # remove last dot from path (if exists) if path.endswith("."): @@ -460,32 +447,6 @@ class NsLcm(LcmBase): additional_params = vdur.get("additionalParams") return parse_yaml_strings(additional_params) - def vnfd2RO(self, vnfd, new_id=None, additionalParams=None, nsrId=None): - """ - Converts creates a new vnfd descriptor for RO base on input OSM IM vnfd - :param vnfd: input vnfd - :param new_id: overrides vnf id if provided - :param additionalParams: Instantiation params for VNFs provided - :param nsrId: Id of the NSR - :return: copy of vnfd - """ - vnfd_RO = deepcopy(vnfd) - # remove unused by RO configuration, monitoring, scaling and internal keys - vnfd_RO.pop("_id", None) - vnfd_RO.pop("_admin", None) - vnfd_RO.pop("monitoring-param", None) - vnfd_RO.pop("scaling-group-descriptor", None) - vnfd_RO.pop("kdu", None) - vnfd_RO.pop("k8s-cluster", None) - if new_id: - vnfd_RO["id"] = new_id - - # parse cloud-init or cloud-init-file with the provided variables using Jinja2 - for vdu in get_iterable(vnfd_RO, "vdu"): - vdu.pop("cloud-init-file", None) - vdu.pop("cloud-init", None) - return vnfd_RO - @staticmethod def ip_profile_2_RO(ip_profile): RO_ip_profile = deepcopy(ip_profile) @@ -504,31 +465,6 @@ class NsLcm(LcmBase): RO_ip_profile["dhcp"] = RO_ip_profile.pop("dhcp-params") return RO_ip_profile - def _get_ro_vim_id_for_vim_account(self, vim_account): - db_vim = self.db.get_one("vim_accounts", {"_id": vim_account}) - if db_vim["_admin"]["operationalState"] != "ENABLED": - raise LcmException( - "VIM={} is not available. operationalState={}".format( - vim_account, db_vim["_admin"]["operationalState"] - ) - ) - RO_vim_id = db_vim["_admin"]["deployed"]["RO"] - return RO_vim_id - - def get_ro_wim_id_for_wim_account(self, wim_account): - if isinstance(wim_account, str): - db_wim = self.db.get_one("wim_accounts", {"_id": wim_account}) - if db_wim["_admin"]["operationalState"] != "ENABLED": - raise LcmException( - "WIM={} is not available. operationalState={}".format( - wim_account, db_wim["_admin"]["operationalState"] - ) - ) - RO_wim_id = db_wim["_admin"]["deployed"]["RO-account"] - return RO_wim_id - else: - return wim_account - def scale_vnfr(self, db_vnfr, vdu_create=None, vdu_delete=None, mark_delete=False): db_vdu_push_list = [] template_vdur = [] @@ -681,103 +617,6 @@ class NsLcm(LcmBase): except DbException as e: self.logger.error("Cannot update vnf. {}".format(e)) - def ns_update_vnfr(self, db_vnfrs, nsr_desc_RO): - """ - Updates database vnfr with the RO info, e.g. ip_address, vim_id... Descriptor db_vnfrs is also updated - :param db_vnfrs: dictionary with member-vnf-index: vnfr-content - :param nsr_desc_RO: nsr descriptor from RO - :return: Nothing, LcmException is raised on errors - """ - for vnf_index, db_vnfr in db_vnfrs.items(): - for vnf_RO in nsr_desc_RO["vnfs"]: - if vnf_RO["member_vnf_index"] != vnf_index: - continue - vnfr_update = {} - if vnf_RO.get("ip_address"): - db_vnfr["ip-address"] = vnfr_update["ip-address"] = vnf_RO[ - "ip_address" - ].split(";")[0] - elif not db_vnfr.get("ip-address"): - if db_vnfr.get("vdur"): # if not VDUs, there is not ip_address - raise LcmExceptionNoMgmtIP( - "ns member_vnf_index '{}' has no IP address".format( - vnf_index - ) - ) - - for vdu_index, vdur in enumerate(get_iterable(db_vnfr, "vdur")): - vdur_RO_count_index = 0 - if vdur.get("pdu-type"): - continue - for vdur_RO in get_iterable(vnf_RO, "vms"): - if vdur["vdu-id-ref"] != vdur_RO["vdu_osm_id"]: - continue - if vdur["count-index"] != vdur_RO_count_index: - vdur_RO_count_index += 1 - continue - vdur["vim-id"] = vdur_RO.get("vim_vm_id") - if vdur_RO.get("ip_address"): - vdur["ip-address"] = vdur_RO["ip_address"].split(";")[0] - else: - vdur["ip-address"] = None - vdur["vdu-id-ref"] = vdur_RO.get("vdu_osm_id") - vdur["name"] = vdur_RO.get("vim_name") - vdur["status"] = vdur_RO.get("status") - vdur["status-detailed"] = vdur_RO.get("error_msg") - for ifacer in get_iterable(vdur, "interfaces"): - for interface_RO in get_iterable(vdur_RO, "interfaces"): - if ifacer["name"] == interface_RO.get("internal_name"): - ifacer["ip-address"] = interface_RO.get( - "ip_address" - ) - ifacer["mac-address"] = interface_RO.get( - "mac_address" - ) - break - else: - raise LcmException( - "ns_update_vnfr: Not found member_vnf_index={} vdur={} interface={} " - "from VIM info".format( - vnf_index, vdur["vdu-id-ref"], ifacer["name"] - ) - ) - vnfr_update["vdur.{}".format(vdu_index)] = vdur - break - else: - raise LcmException( - "ns_update_vnfr: Not found member_vnf_index={} vdur={} count_index={} from " - "VIM info".format( - vnf_index, vdur["vdu-id-ref"], vdur["count-index"] - ) - ) - - for vld_index, vld in enumerate(get_iterable(db_vnfr, "vld")): - for net_RO in get_iterable(nsr_desc_RO, "nets"): - if vld["id"] != net_RO.get("vnf_net_osm_id"): - continue - vld["vim-id"] = net_RO.get("vim_net_id") - vld["name"] = net_RO.get("vim_name") - vld["status"] = net_RO.get("status") - vld["status-detailed"] = net_RO.get("error_msg") - vnfr_update["vld.{}".format(vld_index)] = vld - break - else: - raise LcmException( - "ns_update_vnfr: Not found member_vnf_index={} vld={} from VIM info".format( - vnf_index, vld["id"] - ) - ) - - self.update_db_2("vnfrs", db_vnfr["_id"], vnfr_update) - break - - else: - raise LcmException( - "ns_update_vnfr: Not found member_vnf_index={} from VIM info".format( - vnf_index - ) - ) - def _get_ns_config_info(self, nsr_id): """ Generates a mapping between vnf,vdu elements and the N2VC id @@ -899,7 +738,7 @@ class NsLcm(LcmBase): ) vdur = next((vdur for vdur in target_vnf.get("vdur", ())), None) if not vdur: - return + continue for a_index, a_vld in enumerate(target["ns"]["vld"]): target_vld = find_in_list( get_iterable(vdur, "interfaces"), @@ -967,6 +806,10 @@ class NsLcm(LcmBase): image["vim_info"] = {} for flavor in target["flavor"]: flavor["vim_info"] = {} + if db_nsr.get("shared-volumes"): + target["shared-volumes"] = deepcopy(db_nsr["shared-volumes"]) + for shared_volumes in target["shared-volumes"]: + shared_volumes["vim_info"] = {} if db_nsr.get("affinity-or-anti-affinity-group"): target["affinity-or-anti-affinity-group"] = deepcopy( db_nsr["affinity-or-anti-affinity-group"] @@ -1237,6 +1080,15 @@ class NsLcm(LcmBase): if target_vim not in ns_ags["vim_info"]: ns_ags["vim_info"][target_vim] = {} + # shared-volumes + if vdur.get("shared-volumes-id"): + for sv_id in vdur["shared-volumes-id"]: + ns_sv = find_in_list( + target["shared-volumes"], lambda sv: sv_id in sv["id"] + ) + if ns_sv: + ns_sv["vim_info"][target_vim] = {} + vdur["vim_info"] = {target_vim: {}} # instantiation parameters if vnf_params: @@ -1250,6 +1102,9 @@ class NsLcm(LcmBase): vdu_instantiation_params, vdud ) vdur["additionalParams"]["OSM"]["vdu_volumes"] = vdu_volumes + vdur["additionalParams"]["OSM"][ + "vim_flavor_id" + ] = vdu_instantiation_params.get("vim-flavor-id") vdur_list.append(vdur) target_vnf["vdur"] = vdur_list target["vnf"].append(target_vnf) @@ -1315,7 +1170,7 @@ class NsLcm(LcmBase): db_nsr_update["detailed-status"] = " ".join(stage) self.update_db_2("nsrs", nsr_id, db_nsr_update) self._write_op_status(nslcmop_id, stage) - await asyncio.sleep(15, loop=self.loop) + await asyncio.sleep(15) else: # timeout_ns_deploy raise NgRoException("Timeout waiting ns to deploy") @@ -1501,7 +1356,7 @@ class NsLcm(LcmBase): "target KDU={} is in error state".format(kdu_name) ) - await asyncio.sleep(10, loop=self.loop) + await asyncio.sleep(10) nb_tries += 1 raise LcmException("Timeout waiting KDU={} instantiated".format(kdu_name)) @@ -1532,7 +1387,7 @@ class NsLcm(LcmBase): "Not found _admin.deployed.RO.nsr_id for nsr_id: {}".format(nsr_id) ) - await asyncio.sleep(10, loop=self.loop) + await asyncio.sleep(10) # get ip address if not target_vdu_id: @@ -1818,7 +1673,7 @@ class NsLcm(LcmBase): ee_id, credentials = await self.vca_map[ vca_type ].create_execution_environment( - namespace=namespace, + namespace=nsr_id, reuse_ee_id=ee_id, db_dict=db_dict, config=osm_config, @@ -2293,9 +2148,7 @@ class NsLcm(LcmBase): self.logger.debug( logging_text + "Invoke and wait for placement optimization" ) - await self.msg.aiowrite( - "pla", "get_placement", {"nslcmopId": nslcmop_id}, loop=self.loop - ) + await self.msg.aiowrite("pla", "get_placement", {"nslcmopId": nslcmop_id}) db_poll_interval = 5 wait = db_poll_interval * 10 pla_result = None @@ -2324,6 +2177,299 @@ class NsLcm(LcmBase): vnfr["vim-account-id"] = pla_vnf["vimAccountId"] return modified + def _gather_vnfr_healing_alerts(self, vnfr, vnfd): + alerts = [] + nsr_id = vnfr["nsr-id-ref"] + df = vnfd.get("df", [{}])[0] + # Checking for auto-healing configuration + if "healing-aspect" in df: + healing_aspects = df["healing-aspect"] + for healing in healing_aspects: + for healing_policy in healing.get("healing-policy", ()): + vdu_id = healing_policy["vdu-id"] + vdur = next( + (vdur for vdur in vnfr["vdur"] if vdu_id == vdur["vdu-id-ref"]), + {}, + ) + if not vdur: + continue + metric_name = "vm_status" + vdu_name = vdur.get("name") + vnf_member_index = vnfr["member-vnf-index-ref"] + uuid = str(uuid4()) + name = f"healing_{uuid}" + action = healing_policy + # action_on_recovery = healing.get("action-on-recovery") + # cooldown_time = healing.get("cooldown-time") + # day1 = healing.get("day1") + alert = { + "uuid": uuid, + "name": name, + "metric": metric_name, + "tags": { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_name": vdu_name, + }, + "alarm_status": "ok", + "action_type": "healing", + "action": action, + } + alerts.append(alert) + return alerts + + def _gather_vnfr_scaling_alerts(self, vnfr, vnfd): + alerts = [] + nsr_id = vnfr["nsr-id-ref"] + df = vnfd.get("df", [{}])[0] + # Checking for auto-scaling configuration + if "scaling-aspect" in df: + scaling_aspects = df["scaling-aspect"] + all_vnfd_monitoring_params = {} + for ivld in vnfd.get("int-virtual-link-desc", ()): + for mp in ivld.get("monitoring-parameters", ()): + all_vnfd_monitoring_params[mp.get("id")] = mp + for vdu in vnfd.get("vdu", ()): + for mp in vdu.get("monitoring-parameter", ()): + all_vnfd_monitoring_params[mp.get("id")] = mp + for df in vnfd.get("df", ()): + for mp in df.get("monitoring-parameter", ()): + all_vnfd_monitoring_params[mp.get("id")] = mp + for scaling_aspect in scaling_aspects: + scaling_group_name = scaling_aspect.get("name", "") + # Get monitored VDUs + all_monitored_vdus = set() + for delta in scaling_aspect.get("aspect-delta-details", {}).get( + "deltas", () + ): + for vdu_delta in delta.get("vdu-delta", ()): + all_monitored_vdus.add(vdu_delta.get("id")) + monitored_vdurs = list( + filter( + lambda vdur: vdur["vdu-id-ref"] in all_monitored_vdus, + vnfr["vdur"], + ) + ) + if not monitored_vdurs: + self.logger.error( + "Scaling criteria is referring to a vnf-monitoring-param that does not contain a reference to a vdu or vnf metric" + ) + continue + for scaling_policy in scaling_aspect.get("scaling-policy", ()): + if scaling_policy["scaling-type"] != "automatic": + continue + threshold_time = scaling_policy.get("threshold-time", "1") + cooldown_time = scaling_policy.get("cooldown-time", "0") + for scaling_criteria in scaling_policy["scaling-criteria"]: + monitoring_param_ref = scaling_criteria.get( + "vnf-monitoring-param-ref" + ) + vnf_monitoring_param = all_vnfd_monitoring_params[ + monitoring_param_ref + ] + for vdur in monitored_vdurs: + vdu_id = vdur["vdu-id-ref"] + metric_name = vnf_monitoring_param.get("performance-metric") + metric_name = f"osm_{metric_name}" + vnf_member_index = vnfr["member-vnf-index-ref"] + scalein_threshold = scaling_criteria.get( + "scale-in-threshold" + ) + scaleout_threshold = scaling_criteria.get( + "scale-out-threshold" + ) + # Looking for min/max-number-of-instances + instances_min_number = 1 + instances_max_number = 1 + vdu_profile = df["vdu-profile"] + if vdu_profile: + profile = next( + item for item in vdu_profile if item["id"] == vdu_id + ) + instances_min_number = profile.get( + "min-number-of-instances", 1 + ) + instances_max_number = profile.get( + "max-number-of-instances", 1 + ) + + if scalein_threshold: + uuid = str(uuid4()) + name = f"scalein_{uuid}" + operation = scaling_criteria[ + "scale-in-relational-operation" + ] + rel_operator = self.rel_operation_types.get( + operation, "<=" + ) + metric_selector = f'{metric_name}{{ns_id="{nsr_id}", vnf_member_index="{vnf_member_index}", vdu_id="{vdu_id}"}}' + expression = f"(count ({metric_selector}) > {instances_min_number}) and (avg({metric_selector}) {rel_operator} {scalein_threshold})" + labels = { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_id": vdu_id, + } + prom_cfg = { + "alert": name, + "expr": expression, + "for": str(threshold_time) + "m", + "labels": labels, + } + action = scaling_policy + action = { + "scaling-group": scaling_group_name, + "cooldown-time": cooldown_time, + } + alert = { + "uuid": uuid, + "name": name, + "metric": metric_name, + "tags": { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_id": vdu_id, + }, + "alarm_status": "ok", + "action_type": "scale_in", + "action": action, + "prometheus_config": prom_cfg, + } + alerts.append(alert) + + if scaleout_threshold: + uuid = str(uuid4()) + name = f"scaleout_{uuid}" + operation = scaling_criteria[ + "scale-out-relational-operation" + ] + rel_operator = self.rel_operation_types.get( + operation, "<=" + ) + metric_selector = f'{metric_name}{{ns_id="{nsr_id}", vnf_member_index="{vnf_member_index}", vdu_id="{vdu_id}"}}' + expression = f"(count ({metric_selector}) < {instances_max_number}) and (avg({metric_selector}) {rel_operator} {scaleout_threshold})" + labels = { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_id": vdu_id, + } + prom_cfg = { + "alert": name, + "expr": expression, + "for": str(threshold_time) + "m", + "labels": labels, + } + action = scaling_policy + action = { + "scaling-group": scaling_group_name, + "cooldown-time": cooldown_time, + } + alert = { + "uuid": uuid, + "name": name, + "metric": metric_name, + "tags": { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_id": vdu_id, + }, + "alarm_status": "ok", + "action_type": "scale_out", + "action": action, + "prometheus_config": prom_cfg, + } + alerts.append(alert) + return alerts + + def _gather_vnfr_alarm_alerts(self, vnfr, vnfd): + alerts = [] + nsr_id = vnfr["nsr-id-ref"] + vnf_member_index = vnfr["member-vnf-index-ref"] + + # Checking for VNF alarm configuration + for vdur in vnfr["vdur"]: + vdu_id = vdur["vdu-id-ref"] + vdu = next(filter(lambda vdu: vdu["id"] == vdu_id, vnfd["vdu"])) + if "alarm" in vdu: + # Get VDU monitoring params, since alerts are based on them + vdu_monitoring_params = {} + for mp in vdu.get("monitoring-parameter", []): + vdu_monitoring_params[mp.get("id")] = mp + if not vdu_monitoring_params: + self.logger.error( + "VDU alarm refers to a VDU monitoring param, but there are no VDU monitoring params in the VDU" + ) + continue + # Get alarms in the VDU + alarm_descriptors = vdu["alarm"] + # Create VDU alarms for each alarm in the VDU + for alarm_descriptor in alarm_descriptors: + # Check that the VDU alarm refers to a proper monitoring param + alarm_monitoring_param = alarm_descriptor.get( + "vnf-monitoring-param-ref", "" + ) + vdu_specific_monitoring_param = vdu_monitoring_params.get( + alarm_monitoring_param, {} + ) + if not vdu_specific_monitoring_param: + self.logger.error( + "VDU alarm refers to a VDU monitoring param not present in the VDU" + ) + continue + metric_name = vdu_specific_monitoring_param.get( + "performance-metric" + ) + if not metric_name: + self.logger.error( + "VDU alarm refers to a VDU monitoring param that has no associated performance-metric" + ) + continue + # Set params of the alarm to be created in Prometheus + metric_name = f"osm_{metric_name}" + metric_threshold = alarm_descriptor.get("value") + uuid = str(uuid4()) + alert_name = f"vdu_alarm_{uuid}" + operation = alarm_descriptor["operation"] + rel_operator = self.rel_operation_types.get(operation, "<=") + metric_selector = f'{metric_name}{{ns_id="{nsr_id}", vnf_member_index="{vnf_member_index}", vdu_id="{vdu_id}"}}' + expression = f"{metric_selector} {rel_operator} {metric_threshold}" + labels = { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_id": vdu_id, + "vdu_name": "{{ $labels.vdu_name }}", + } + prom_cfg = { + "alert": alert_name, + "expr": expression, + "for": "1m", # default value. Ideally, this should be related to an IM param, but there is not such param + "labels": labels, + } + alarm_action = dict() + for action_type in ["ok", "insufficient-data", "alarm"]: + if ( + "actions" in alarm_descriptor + and action_type in alarm_descriptor["actions"] + ): + alarm_action[action_type] = alarm_descriptor["actions"][ + action_type + ] + alert = { + "uuid": uuid, + "name": alert_name, + "metric": metric_name, + "tags": { + "ns_id": nsr_id, + "vnf_member_index": vnf_member_index, + "vdu_id": vdu_id, + }, + "alarm_status": "ok", + "action_type": "vdu_alarm", + "action": alarm_action, + "prometheus_config": prom_cfg, + } + alerts.append(alert) + return alerts + def update_nsrs_with_pla_result(self, params): try: nslcmop_id = deep_get(params, ("placement", "nslcmopId")) @@ -2533,13 +2679,16 @@ class NsLcm(LcmBase): # create namespace and certificate if any helm based EE is present in the NS if check_helm_ee_in_ns(db_vnfds): - # TODO: create EE namespace + await self.vca_map["helm-v3"].setup_ns_namespace( + name=nsr_id, + ) # create TLS certificates await self.vca_map["helm-v3"].create_tls_certificate( - secret_name="ee-tls-{}".format(nsr_id), + secret_name=self.EE_TLS_NAME, dns_prefix="*", nsr_id=nsr_id, usage="server auth", + namespace=nsr_id, ) nsi_id = None # TODO put nsi_id when this nsr belongs to a NSI @@ -2675,6 +2824,65 @@ class NsLcm(LcmBase): stage=stage, ) + # Check if each vnf has exporter for metric collection if so update prometheus job records + if "exporters-endpoints" in vnfd.get("df")[0]: + exporter_config = vnfd.get("df")[0].get("exporters-endpoints") + self.logger.debug("exporter config :{}".format(exporter_config)) + artifact_path = "{}/{}/{}".format( + base_folder["folder"], + base_folder["pkg-dir"], + "exporter-endpoint", + ) + ee_id = None + ee_config_descriptor = exporter_config + vnfr_id = db_vnfr["id"] + rw_mgmt_ip = await self.wait_vm_up_insert_key_ro( + logging_text, + nsr_id, + vnfr_id, + vdu_id=None, + vdu_index=None, + user=None, + pub_key=None, + ) + self.logger.debug("rw_mgmt_ip:{}".format(rw_mgmt_ip)) + self.logger.debug("Artifact_path:{}".format(artifact_path)) + db_vnfr = self.db.get_one("vnfrs", {"_id": vnfr_id}) + vdu_id_for_prom = None + vdu_index_for_prom = None + for x in get_iterable(db_vnfr, "vdur"): + vdu_id_for_prom = x.get("vdu-id-ref") + vdu_index_for_prom = x.get("count-index") + prometheus_jobs = await self.extract_prometheus_scrape_jobs( + ee_id=ee_id, + artifact_path=artifact_path, + ee_config_descriptor=ee_config_descriptor, + vnfr_id=vnfr_id, + nsr_id=nsr_id, + target_ip=rw_mgmt_ip, + element_type="VDU", + vdu_id=vdu_id_for_prom, + vdu_index=vdu_index_for_prom, + ) + + self.logger.debug("Prometheus job:{}".format(prometheus_jobs)) + if prometheus_jobs: + db_nsr_update["_admin.deployed.prometheus_jobs"] = prometheus_jobs + self.update_db_2( + "nsrs", + nsr_id, + db_nsr_update, + ) + + for job in prometheus_jobs: + self.db.set_one( + "prometheus_jobs", + {"job_name": job["job_name"]}, + job, + upsert=True, + fail_on_empty=False, + ) + # Check if this NS has a charm configuration descriptor_config = nsd.get("ns-configuration") if descriptor_config and descriptor_config.get("juju"): @@ -2798,7 +3006,27 @@ class NsLcm(LcmBase): db_nsr_update["detailed-status"] = "Done" db_nslcmop_update["detailed-status"] = "Done" nslcmop_operation_state = "COMPLETED" - + # Gather auto-healing and auto-scaling alerts for each vnfr + healing_alerts = [] + scaling_alerts = [] + for vnfr in self.db.get_list("vnfrs", {"nsr-id-ref": nsr_id}): + vnfd = next( + (sub for sub in db_vnfds if sub["_id"] == vnfr["vnfd-id"]), None + ) + healing_alerts = self._gather_vnfr_healing_alerts(vnfr, vnfd) + for alert in healing_alerts: + self.logger.info(f"Storing healing alert in MongoDB: {alert}") + self.db.create("alerts", alert) + + scaling_alerts = self._gather_vnfr_scaling_alerts(vnfr, vnfd) + for alert in scaling_alerts: + self.logger.info(f"Storing scaling alert in MongoDB: {alert}") + self.db.create("alerts", alert) + + alarm_alerts = self._gather_vnfr_alarm_alerts(vnfr, vnfd) + for alert in alarm_alerts: + self.logger.info(f"Storing VNF alarm alert in MongoDB: {alert}") + self.db.create("alerts", alert) if db_nsr: self._write_ns_status( nsr_id=nsr_id, @@ -2826,8 +3054,13 @@ class NsLcm(LcmBase): "nsr_id": nsr_id, "nslcmop_id": nslcmop_id, "operationState": nslcmop_operation_state, + "startTime": db_nslcmop["startTime"], + "links": db_nslcmop["links"], + "operationParams": { + "nsInstanceId": nsr_id, + "nsdId": db_nsr["nsd-id"], + }, }, - loop=self.loop, ) except Exception as e: self.logger.error( @@ -3839,41 +4072,6 @@ class NsLcm(LcmBase): member_vnf_index or "", vdu_id or "" ) - @staticmethod - def _create_nslcmop(nsr_id, operation, params): - """ - Creates a ns-lcm-opp content to be stored at database. - :param nsr_id: internal id of the instance - :param operation: instantiate, terminate, scale, action, ... - :param params: user parameters for the operation - :return: dictionary following SOL005 format - """ - # Raise exception if invalid arguments - if not (nsr_id and operation and params): - raise LcmException( - "Parameters 'nsr_id', 'operation' and 'params' needed to create primitive not provided" - ) - now = time() - _id = str(uuid4()) - nslcmop = { - "id": _id, - "_id": _id, - # COMPLETED,PARTIALLY_COMPLETED,FAILED_TEMP,FAILED,ROLLING_BACK,ROLLED_BACK - "operationState": "PROCESSING", - "statusEnteredTime": now, - "nsInstanceId": nsr_id, - "lcmOperationType": operation, - "startTime": now, - "isAutomaticInvocation": False, - "operationParams": params, - "isCancelPending": False, - "links": { - "self": "/osm/nslcm/v1/ns_lcm_op_occs/" + _id, - "nsInstance": "/osm/nslcm/v1/ns_instances/" + nsr_id, - }, - } - return nslcmop - def _format_additional_params(self, params): params = params or {} for key, value in params.items(): @@ -4069,12 +4267,6 @@ class NsLcm(LcmBase): # Function to return execution_environment id - def _get_ee_id(self, vnf_index, vdu_id, vca_deployed_list): - # TODO vdu_index_count - for vca in vca_deployed_list: - if vca["member-vnf-index"] == vnf_index and vca["vdu_id"] == vdu_id: - return vca["ee_id"] - async def destroy_N2VC( self, logging_text, @@ -4355,9 +4547,12 @@ class NsLcm(LcmBase): # Delete Namespace and Certificates if necessary if check_helm_ee_in_ns(list(db_vnfds_from_member_index.values())): await self.vca_map["helm-v3"].delete_tls_certificate( - certificate_name=db_nslcmop["nsInstanceId"], + namespace=db_nslcmop["nsInstanceId"], + certificate_name=self.EE_TLS_NAME, + ) + await self.vca_map["helm-v3"].delete_namespace( + namespace=db_nslcmop["nsInstanceId"], ) - # TODO: Delete namespace # Delete from k8scluster stage[1] = "Deleting KDUs." @@ -4514,12 +4709,13 @@ class NsLcm(LcmBase): "operationState": nslcmop_operation_state, "autoremove": autoremove, }, - loop=self.loop, ) except Exception as e: self.logger.error( logging_text + "kafka_write notification Exception {}".format(e) ) + self.logger.debug(f"Deleting alerts: ns_id={nsr_id}") + self.db.del_list("alerts", {"tags.ns_id": nsr_id}) self.logger.debug(logging_text + "Exit") self.lcm_tasks.remove("ns", nsr_id, nslcmop_id, "ns_terminate") @@ -4770,7 +4966,7 @@ class NsLcm(LcmBase): ) ) # wait and retry - await asyncio.sleep(retries_interval, loop=self.loop) + await asyncio.sleep(retries_interval) else: if isinstance(e, asyncio.TimeoutError): e = N2VCException( @@ -5010,9 +5206,10 @@ class NsLcm(LcmBase): del desc_params["kdu_model"] else: kdu_model = kdu.get("kdu-model") - parts = kdu_model.split(sep=":") - if len(parts) == 2: - kdu_model = parts[0] + if kdu_model.count("/") < 2: # helm chart is not embedded + parts = kdu_model.split(sep=":") + if len(parts) == 2: + kdu_model = parts[0] if desc_params.get("kdu_atomic_upgrade"): atomic_upgrade = desc_params.get( "kdu_atomic_upgrade" @@ -5181,7 +5378,6 @@ class NsLcm(LcmBase): "nslcmop_id": nslcmop_id, "operationState": nslcmop_operation_state, }, - loop=self.loop, ) except Exception as e: self.logger.error( @@ -6012,7 +6208,7 @@ class NsLcm(LcmBase): and member_vnf_index ): msg.update({"vnf_member_index": member_vnf_index}) - await self.msg.aiowrite("ns", change_type, msg, loop=self.loop) + await self.msg.aiowrite("ns", change_type, msg) except Exception as e: self.logger.error( logging_text + "kafka_write notification Exception {}".format(e) @@ -7045,7 +7241,7 @@ class NsLcm(LcmBase): "nslcmop_id": nslcmop_id, "operationState": nslcmop_operation_state, } - await self.msg.aiowrite("ns", "scaled", msg, loop=self.loop) + await self.msg.aiowrite("ns", "scaled", msg) except Exception as e: self.logger.error( logging_text + "kafka_write notification Exception {}".format(e) @@ -7265,6 +7461,8 @@ class NsLcm(LcmBase): ) if not job_file: return + self.logger.debug("Artifact path{}".format(artifact_path)) + self.logger.debug("job file{}".format(job_file)) with self.fs.file_open((artifact_path, job_file), "r") as f: job_data = f.read() @@ -7303,7 +7501,7 @@ class NsLcm(LcmBase): kdur_name = kdur.get("name") break - await asyncio.sleep(10, loop=self.loop) + await asyncio.sleep(10) else: if vdu_id and vdu_index is not None: raise LcmException( @@ -7314,22 +7512,35 @@ class NsLcm(LcmBase): f"Timeout waiting KDU with name={kdu_name} and index={kdu_index} to be intantiated" ) - # TODO get_service - _, _, service = ee_id.partition(".") # remove prefix "namespace." - host_name = "{}-{}".format(service, ee_config_descriptor["metric-service"]) - host_port = "80" - vnfr_id = vnfr_id.replace("-", "") - variables = { - "JOB_NAME": vnfr_id, - "TARGET_IP": target_ip, - "EXPORTER_POD_IP": host_name, - "EXPORTER_POD_PORT": host_port, - "NSR_ID": nsr_id, - "VNF_MEMBER_INDEX": vnf_member_index, - "VDUR_NAME": vdur_name, - "KDUR_NAME": kdur_name, - "ELEMENT_TYPE": element_type, - } + if ee_id is not None: + _, namespace, helm_id = get_ee_id_parts( + ee_id + ) # get namespace and EE gRPC service name + host_name = f'{helm_id}-{ee_config_descriptor["metric-service"]}.{namespace}.svc' # svc_name.namespace.svc + host_port = "80" + vnfr_id = vnfr_id.replace("-", "") + variables = { + "JOB_NAME": vnfr_id, + "TARGET_IP": target_ip, + "EXPORTER_POD_IP": host_name, + "EXPORTER_POD_PORT": host_port, + "NSR_ID": nsr_id, + "VNF_MEMBER_INDEX": vnf_member_index, + "VDUR_NAME": vdur_name, + "KDUR_NAME": kdur_name, + "ELEMENT_TYPE": element_type, + } + else: + metric_path = ee_config_descriptor["metric-path"] + target_port = ee_config_descriptor["metric-port"] + vnfr_id = vnfr_id.replace("-", "") + variables = { + "JOB_NAME": vnfr_id, + "TARGET_IP": target_ip, + "TARGET_PORT": target_port, + "METRIC_PATH": metric_path, + } + job_list = parse_job(job_data, variables) # ensure job_name is using the vnfr_id. Adding the metadata nsr_id for job in job_list: @@ -7337,7 +7548,7 @@ class NsLcm(LcmBase): not isinstance(job.get("job_name"), str) or vnfr_id not in job["job_name"] ): - job["job_name"] = vnfr_id + "_" + str(randint(1, 10000)) + job["job_name"] = vnfr_id + "_" + str(SystemRandom().randint(1, 10000)) job["nsr_id"] = nsr_id job["vnfr_id"] = vnfr_id return job_list @@ -7429,28 +7640,6 @@ class NsLcm(LcmBase): ) return "FAILED", "Error in operate VNF {}".format(exc) - def get_vca_cloud_and_credentials(self, vim_account_id: str) -> (str, str): - """ - Get VCA Cloud and VCA Cloud Credentials for the VIM account - - :param: vim_account_id: VIM Account ID - - :return: (cloud_name, cloud_credential) - """ - config = VimAccountDB.get_vim_account_with_id(vim_account_id).get("config", {}) - return config.get("vca_cloud"), config.get("vca_cloud_credential") - - def get_vca_k8s_cloud_and_credentials(self, vim_account_id: str) -> (str, str): - """ - Get VCA K8s Cloud and VCA K8s Cloud Credentials for the VIM account - - :param: vim_account_id: VIM Account ID - - :return: (cloud_name, cloud_credential) - """ - config = VimAccountDB.get_vim_account_with_id(vim_account_id).get("config", {}) - return config.get("vca_k8s_cloud"), config.get("vca_k8s_cloud_credential") - async def migrate(self, nsr_id, nslcmop_id): """ Migrate VNFs and VDUs instances in a NS @@ -7546,7 +7735,7 @@ class NsLcm(LcmBase): "nslcmop_id": nslcmop_id, "operationState": nslcmop_operation_state, } - await self.msg.aiowrite("ns", "migrated", msg, loop=self.loop) + await self.msg.aiowrite("ns", "migrated", msg) except Exception as e: self.logger.error( logging_text + "kafka_write notification Exception {}".format(e) @@ -7670,9 +7859,9 @@ class NsLcm(LcmBase): for target_vdu in target_vdu_list: deploy_params_vdu = target_vdu # Set run-day1 vnf level value if not vdu level value exists - if not deploy_params_vdu.get("run-day1") and target_vnf[ - "additionalParams" - ].get("run-day1"): + if not deploy_params_vdu.get("run-day1") and target_vnf.get( + "additionalParams", {} + ).get("run-day1"): deploy_params_vdu["run-day1"] = target_vnf[ "additionalParams" ].get("run-day1") @@ -7832,7 +8021,7 @@ class NsLcm(LcmBase): "nslcmop_id": nslcmop_id, "operationState": nslcmop_operation_state, } - await self.msg.aiowrite("ns", "healed", msg, loop=self.loop) + await self.msg.aiowrite("ns", "healed", msg) except Exception as e: self.logger.error( logging_text + "kafka_write notification Exception {}".format(e) @@ -8513,7 +8702,7 @@ class NsLcm(LcmBase): self.logger.debug("Wait Heal RO > {}".format(operational_status_ro)) if operational_status_ro != "healing": break - await asyncio.sleep(15, loop=self.loop) + await asyncio.sleep(15) else: # timeout_ns_deploy raise NgRoException("Timeout waiting ns to deploy") @@ -8611,7 +8800,7 @@ class NsLcm(LcmBase): "nslcmop_id": nslcmop_id, "operationState": nslcmop_operation_state, } - await self.msg.aiowrite("ns", "verticalscaled", msg, loop=self.loop) + await self.msg.aiowrite("ns", "verticalscaled", msg) except Exception as e: self.logger.error( logging_text + "kafka_write notification Exception {}".format(e)