check_juju_bundle_existence,
get_charm_artifact_path,
get_ee_id_parts,
+ vld_to_ro_ip_profile,
)
from osm_lcm.data_utils.nsd import (
get_ns_configuration_relation_list,
from osm_lcm.data_utils.dict_utils import parse_yaml_strings
from osm_lcm.data_utils.database.vim_account import VimAccountDB
from n2vc.definitions import RelationEndpoint
-from n2vc.k8s_helm_conn import K8sHelmConnector
from n2vc.k8s_helm3_conn import K8sHelm3Connector
from n2vc.k8s_juju_conn import K8sJujuConnector
from time import time
from uuid import uuid4
-from random import randint
+from random import SystemRandom
__author__ = "Alfonso Tierno <alfonso.tiernosepulveda@telefonica.com>"
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',
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
# 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,
self.conn_helm_ee = LCMHelmConn(
log=self.logger,
- loop=self.loop,
vca_config=self.vca_config,
on_update_db=self._on_update_n2vc_db,
)
- self.k8sclusterhelm2 = K8sHelmConnector(
- kubectl_command=self.vca_config.kubectlpath,
- helm_command=self.vca_config.helmpath,
- log=self.logger,
- on_update_db=None,
- fs=self.fs,
- db=self.db,
- )
-
self.k8sclusterhelm3 = K8sHelm3Connector(
kubectl_command=self.vca_config.kubectlpath,
helm_command=self.vca_config.helm3path,
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,
)
self.k8scluster_map = {
- "helm-chart": self.k8sclusterhelm2,
"helm-chart-v3": self.k8sclusterhelm3,
"chart": self.k8sclusterhelm3,
"juju-bundle": self.k8sclusterjuju,
}
# 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,
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:
)
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("."):
path = path[:-1]
# self.logger.debug('_on_update_n2vc_db(table={}, filter={}, path={}, updated_data={}'
# .format(table, filter, path, updated_data))
try:
-
nsr_id = filter.get("_id")
# read ns record from database
return wim_account
def scale_vnfr(self, db_vnfr, vdu_create=None, vdu_delete=None, mark_delete=False):
-
db_vdu_push_list = []
template_vdur = []
db_update = {"_admin.modified": time()}
start_deploy,
timeout_ns_deploy,
):
-
db_vims = {}
def get_vim_account(vim_account_id):
target_vim, target_vld, vld_params, target_sdn
):
if vld_params.get("ip-profile"):
- target_vld["vim_info"][target_vim]["ip_profile"] = vld_params[
- "ip-profile"
- ]
+ target_vld["vim_info"][target_vim]["ip_profile"] = vld_to_ro_ip_profile(
+ vld_params["ip-profile"]
+ )
if vld_params.get("provider-network"):
target_vld["vim_info"][target_vim]["provider_network"] = vld_params[
"provider-network"
)
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"),
lambda v_vld: v_vld["name"] in (a_vld["name"], a_vld["id"]),
)
if target_vld:
-
if vnf_params.get("vimAccountId") not in a_vld.get(
"vim_info", {}
):
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"]
# check if this network needs SDN assist
if vld.get("pci-interfaces"):
db_vim = get_vim_account(ns_params["vimAccountId"])
- sdnc_id = db_vim["config"].get("sdn-controller")
- if sdnc_id:
- sdn_vld = "nsrs:{}:vld.{}".format(nsr_id, vld["id"])
- target_sdn = "sdn:{}".format(sdnc_id)
- target_vld["vim_info"][target_sdn] = {
- "sdn": True,
- "target_vim": target_vim,
- "vlds": [sdn_vld],
- "type": vld.get("type"),
- }
+ if vim_config := db_vim.get("config"):
+ if sdnc_id := vim_config.get("sdn-controller"):
+ sdn_vld = "nsrs:{}:vld.{}".format(nsr_id, vld["id"])
+ target_sdn = "sdn:{}".format(sdnc_id)
+ target_vld["vim_info"][target_sdn] = {
+ "sdn": True,
+ "target_vim": target_vim,
+ "vlds": [sdn_vld],
+ "type": vld.get("type"),
+ }
nsd_vnf_profiles = get_vnf_profiles(nsd)
for nsd_vnf_profile in nsd_vnf_profiles:
and nsd_vlp.get("virtual-link-protocol-data")
and nsd_vlp["virtual-link-protocol-data"].get("l3-protocol-data")
):
- ip_profile_source_data = nsd_vlp["virtual-link-protocol-data"][
+ vld_params["ip-profile"] = nsd_vlp["virtual-link-protocol-data"][
"l3-protocol-data"
]
- ip_profile_dest_data = {}
- if "ip-version" in ip_profile_source_data:
- ip_profile_dest_data["ip-version"] = ip_profile_source_data[
- "ip-version"
- ]
- if "cidr" in ip_profile_source_data:
- ip_profile_dest_data["subnet-address"] = ip_profile_source_data[
- "cidr"
- ]
- if "gateway-ip" in ip_profile_source_data:
- ip_profile_dest_data["gateway-address"] = ip_profile_source_data[
- "gateway-ip"
- ]
- if "dhcp-enabled" in ip_profile_source_data:
- ip_profile_dest_data["dhcp-params"] = {
- "enabled": ip_profile_source_data["dhcp-enabled"]
- }
- vld_params["ip-profile"] = ip_profile_dest_data
# update vld_params with instantiation params
vld_instantiation_params = find_in_list(
and vnfd_vlp.get("virtual-link-protocol-data")
and vnfd_vlp["virtual-link-protocol-data"].get("l3-protocol-data")
):
- ip_profile_source_data = vnfd_vlp["virtual-link-protocol-data"][
+ vld_params["ip-profile"] = vnfd_vlp["virtual-link-protocol-data"][
"l3-protocol-data"
]
- ip_profile_dest_data = {}
- if "ip-version" in ip_profile_source_data:
- ip_profile_dest_data["ip-version"] = ip_profile_source_data[
- "ip-version"
- ]
- if "cidr" in ip_profile_source_data:
- ip_profile_dest_data["subnet-address"] = ip_profile_source_data[
- "cidr"
- ]
- if "gateway-ip" in ip_profile_source_data:
- ip_profile_dest_data[
- "gateway-address"
- ] = ip_profile_source_data["gateway-ip"]
- if "dhcp-enabled" in ip_profile_source_data:
- ip_profile_dest_data["dhcp-params"] = {
- "enabled": ip_profile_source_data["dhcp-enabled"]
- }
-
- vld_params["ip-profile"] = ip_profile_dest_data
# update vld_params with instantiation params
if vnf_params:
vld_instantiation_params = find_in_list(
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:
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)
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")
}
desc = await self.RO.deploy(nsr_id, target)
action_id = desc["action_id"]
- db_nsr_update["_admin.deployed.RO.nsr_delete_action_id"] = action_id
db_nsr_update["_admin.deployed.RO.nsr_status"] = "DELETING"
self.logger.debug(
logging_text
stage,
operation="termination",
)
-
- db_nsr_update["_admin.deployed.RO.nsr_delete_action_id"] = None
db_nsr_update["_admin.deployed.RO.nsr_status"] = "DELETED"
# delete all nsr
await self.RO.delete(nsr_id)
- except Exception as e:
- if isinstance(e, NgRoException) and e.http_code == 404: # not found
+ except NgRoException as e:
+ if e.http_code == 404: # not found
db_nsr_update["_admin.deployed.RO.nsr_id"] = None
db_nsr_update["_admin.deployed.RO.nsr_status"] = "DELETED"
- db_nsr_update["_admin.deployed.RO.nsr_delete_action_id"] = None
self.logger.debug(
logging_text + "RO_action_id={} already deleted".format(action_id)
)
- elif isinstance(e, NgRoException) and e.http_code == 409: # conflict
+ elif e.http_code == 409: # conflict
failed_detail.append("delete conflict: {}".format(e))
self.logger.debug(
logging_text
logging_text
+ "RO_action_id={} delete error: {}".format(action_id, e)
)
+ except Exception as e:
+ failed_detail.append("delete error: {}".format(e))
+ self.logger.error(
+ logging_text + "RO_action_id={} delete error: {}".format(action_id, e)
+ )
if failed_detail:
stage[2] = "Error deleting from VIM"
"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))
"""
self.logger.debug(logging_text + "Starting wait_vm_up_insert_key_ro")
- ro_nsr_id = None
ip_address = None
- nb_tries = 0
target_vdu_id = None
ro_retries = 0
while True:
-
ro_retries += 1
if ro_retries >= 360: # 1 hour
raise LcmException(
"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:
self.logger.error(logging_text + "Cannot inject ssh-ky to a PDU")
return ip_address
try:
- ro_vm_id = "{}-{}".format(
- db_vnfr["member-vnf-index-ref"], target_vdu_id
- ) # TODO add vdu_index
- if self.ro_config.ng:
- target = {
- "action": {
- "action": "inject_ssh_key",
- "key": pub_key,
- "user": user,
- },
- "vnf": [{"_id": vnfr_id, "vdur": [{"id": vdur["id"]}]}],
- }
- desc = await self.RO.deploy(nsr_id, target)
- action_id = desc["action_id"]
- await self._wait_ng_ro(
- nsr_id, action_id, timeout=600, operation="instantiation"
- )
- break
- else:
- # wait until NS is deployed at RO
- if not ro_nsr_id:
- db_nsrs = self.db.get_one("nsrs", {"_id": nsr_id})
- ro_nsr_id = deep_get(
- db_nsrs, ("_admin", "deployed", "RO", "nsr_id")
- )
- if not ro_nsr_id:
- continue
- result_dict = await self.RO.create_action(
- item="ns",
- item_id_name=ro_nsr_id,
- descriptor={
- "add_public_key": pub_key,
- "vms": [ro_vm_id],
- "user": user,
- },
- )
- # result_dict contains the format {VM-id: {vim_result: 200, description: text}}
- if not result_dict or not isinstance(result_dict, dict):
- raise LcmException(
- "Unknown response from RO when injecting key"
- )
- for result in result_dict.values():
- if result.get("vim_result") == 200:
- break
- else:
- raise ROclient.ROClientException(
- "error injecting key: {}".format(
- result.get("description")
- )
- )
- break
+ target = {
+ "action": {
+ "action": "inject_ssh_key",
+ "key": pub_key,
+ "user": user,
+ },
+ "vnf": [{"_id": vnfr_id, "vdur": [{"id": vdur["id"]}]}],
+ }
+ desc = await self.RO.deploy(nsr_id, target)
+ action_id = desc["action_id"]
+ await self._wait_ng_ro(
+ nsr_id, action_id, timeout=600, operation="instantiation"
+ )
+ break
except NgRoException as e:
raise LcmException(
"Reaching max tries injecting key. Error: {}".format(e)
)
- except ROclient.ROClientException as e:
- if not nb_tries:
- self.logger.debug(
- logging_text
- + "error injecting key: {}. Retrying until {} seconds".format(
- e, 20 * 10
- )
- )
- nb_tries += 1
- if nb_tries >= 20:
- raise LcmException(
- "Reaching max tries injecting key. Error: {}".format(e)
- )
else:
break
}
step = ""
try:
-
element_type = "NS"
element_under_configuration = nsr_id
vca_id = self.get_vca_id(db_vnfr, db_nsr)
# create or register execution environment in VCA
- if vca_type in ("lxc_proxy_charm", "k8s_proxy_charm", "helm", "helm-v3"):
-
+ if vca_type in ("lxc_proxy_charm", "k8s_proxy_charm", "helm-v3"):
self._write_configuration_status(
nsr_id=nsr_id,
vca_index=vca_index,
db_dict=db_dict,
vca_id=vca_id,
)
- elif vca_type == "helm" or vca_type == "helm-v3":
+ elif vca_type == "helm-v3":
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,
# if SSH access is required, then get execution environment SSH public
# if native charm we have waited already to VM be UP
- if vca_type in ("k8s_proxy_charm", "lxc_proxy_charm", "helm", "helm-v3"):
+ if vca_type in ("k8s_proxy_charm", "lxc_proxy_charm", "helm-v3"):
pub_key = None
user = None
# self.logger.debug("get ssh key block")
# TODO register in database that primitive is done
# STEP 7 Configure metrics
- if vca_type == "helm" or vca_type == "helm-v3":
+ if vca_type == "helm-v3":
# TODO: review for those cases where the helm chart is a reference and
# is not part of the NF package
prometheus_jobs = await self.extract_prometheus_scrape_jobs(
vnfr_id=vnfr_id,
nsr_id=nsr_id,
target_ip=rw_mgmt_ip,
+ element_type=element_type,
vnf_member_index=db_vnfr.get("member-vnf-index-ref", ""),
vdu_id=vdu_id,
vdu_index=vdu_index,
element_type: str = None,
other_update: dict = None,
):
-
# self.logger.debug('_write_configuration_status(): vca_index={}, status={}'
# .format(vca_index, status))
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
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"))
# update operation on nslcmops
db_nslcmop_update = {}
+ timeout_ns_deploy = self.timeout.ns_deploy
+
nslcmop_operation_state = None
db_vnfrs = {} # vnf's info indexed by member-index
# n2vc_info = {}
ns_params = db_nslcmop.get("operationParams")
if ns_params and ns_params.get("timeout_ns_deploy"):
timeout_ns_deploy = ns_params["timeout_ns_deploy"]
- else:
- timeout_ns_deploy = self.timeout.ns_deploy
# read from db: ns
stage[1] = "Getting nsr={} from db.".format(nsr_id)
# 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
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"):
stage[1] = stage[2] = ""
except asyncio.CancelledError:
error_list.append("Cancelled")
- # TODO cancel all tasks
+ await self._cancel_pending_tasks(logging_text, tasks_dict_info)
+ await self._wait_for_tasks(
+ logging_text,
+ tasks_dict_info,
+ timeout_ns_deploy,
+ stage,
+ nslcmop_id,
+ nsr_id=nsr_id,
+ )
except Exception as exc:
error_list.append(str(exc))
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,
"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(
cached_vnfds: Dict[str, Any],
) -> List[Relation]:
relations = []
+ if vca.target_element == "ns":
+ self.logger.debug("VCA is a NS charm, not a VNF.")
+ return relations
vnf_profile = get_vnf_profile(nsd, vca.vnf_profile_id)
vnf_profile_id = vnf_profile["id"]
vnfd_id = vnf_profile["vnfd-id"]
vca_index: int,
timeout: int = 3600,
) -> bool:
-
# steps:
# 1. find all relations for this VCA
# 2. wait for other peers related
timeout: int = 600,
vca_id: str = None,
):
-
try:
k8sclustertype = k8s_instance_info["k8scluster-type"]
# Instantiate kdu
vnfr_data.get("_id"),
{"kdur.{}.status".format(kdu_index): "ERROR"},
)
- except Exception:
+ except Exception as error:
# ignore to keep original exception
- pass
+ self.logger.warning(
+ f"An exception occurred while updating DB: {str(error)}"
+ )
# reraise original error
raise
k8scluster_id_2_uuic = {
"helm-chart-v3": {},
- "helm-chart": {},
"juju-bundle": {},
}
# Default version: helm3, if helm-version is v2 assign v2
k8sclustertype = "helm-chart-v3"
self.logger.debug("kdur: {}".format(kdur))
- if (
- kdur.get("helm-version")
- and kdur.get("helm-version") == "v2"
- ):
- k8sclustertype = "helm-chart"
elif kdur.get("juju-bundle"):
kdumodel = kdur["juju-bundle"]
k8sclustertype = "juju-bundle"
kdumodel = self.fs.path + filename
except (asyncio.TimeoutError, asyncio.CancelledError):
raise
- except Exception: # it is not a file
- pass
+ except Exception as e: # it is not a file
+ self.logger.warning(f"An exception occurred: {str(e)}")
k8s_cluster_id = kdur["k8s-cluster"]["id"]
step = "Synchronize repos for k8s cluster '{}'".format(
vca_type = "native_charm"
elif ee_item.get("helm-chart"):
vca_name = ee_item["helm-chart"]
- if ee_item.get("helm-version") and ee_item.get("helm-version") == "v2":
- vca_type = "helm"
- else:
- vca_type = "helm-v3"
+ vca_type = "helm-v3"
else:
self.logger.debug(
logging_text + "skipping non juju neither charm configuration"
# 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"]
+ return vca.get("ee_id")
async def destroy_N2VC(
self,
pass
self._write_all_config_status(db_nsr=db_nsr, status="DELETED")
- async def _terminate_RO(
- self, logging_text, nsr_deployed, nsr_id, nslcmop_id, stage
- ):
- """
- Terminates a deployment from RO
- :param logging_text:
- :param nsr_deployed: db_nsr._admin.deployed
- :param nsr_id:
- :param nslcmop_id:
- :param stage: list of string with the content to write on db_nslcmop.detailed-status.
- this method will update only the index 2, but it will write on database the concatenated content of the list
- :return:
- """
- db_nsr_update = {}
- failed_detail = []
- ro_nsr_id = ro_delete_action = None
- if nsr_deployed and nsr_deployed.get("RO"):
- ro_nsr_id = nsr_deployed["RO"].get("nsr_id")
- ro_delete_action = nsr_deployed["RO"].get("nsr_delete_action_id")
- try:
- if ro_nsr_id:
- stage[2] = "Deleting ns from VIM."
- db_nsr_update["detailed-status"] = " ".join(stage)
- self._write_op_status(nslcmop_id, stage)
- self.logger.debug(logging_text + stage[2])
- self.update_db_2("nsrs", nsr_id, db_nsr_update)
- self._write_op_status(nslcmop_id, stage)
- desc = await self.RO.delete("ns", ro_nsr_id)
- ro_delete_action = desc["action_id"]
- db_nsr_update[
- "_admin.deployed.RO.nsr_delete_action_id"
- ] = ro_delete_action
- db_nsr_update["_admin.deployed.RO.nsr_id"] = None
- db_nsr_update["_admin.deployed.RO.nsr_status"] = "DELETED"
- if ro_delete_action:
- # wait until NS is deleted from VIM
- stage[2] = "Waiting ns deleted from VIM."
- detailed_status_old = None
- self.logger.debug(
- logging_text
- + stage[2]
- + " RO_id={} ro_delete_action={}".format(
- ro_nsr_id, ro_delete_action
- )
- )
- self.update_db_2("nsrs", nsr_id, db_nsr_update)
- self._write_op_status(nslcmop_id, stage)
-
- delete_timeout = 20 * 60 # 20 minutes
- while delete_timeout > 0:
- desc = await self.RO.show(
- "ns",
- item_id_name=ro_nsr_id,
- extra_item="action",
- extra_item_id=ro_delete_action,
- )
-
- # deploymentStatus
- self._on_update_ro_db(nsrs_id=nsr_id, ro_descriptor=desc)
-
- ns_status, ns_status_info = self.RO.check_action_status(desc)
- if ns_status == "ERROR":
- raise ROclient.ROClientException(ns_status_info)
- elif ns_status == "BUILD":
- stage[2] = "Deleting from VIM {}".format(ns_status_info)
- elif ns_status == "ACTIVE":
- db_nsr_update["_admin.deployed.RO.nsr_delete_action_id"] = None
- db_nsr_update["_admin.deployed.RO.nsr_status"] = "DELETED"
- break
- else:
- assert (
- False
- ), "ROclient.check_action_status returns unknown {}".format(
- ns_status
- )
- if stage[2] != detailed_status_old:
- detailed_status_old = stage[2]
- db_nsr_update["detailed-status"] = " ".join(stage)
- self._write_op_status(nslcmop_id, stage)
- self.update_db_2("nsrs", nsr_id, db_nsr_update)
- await asyncio.sleep(5, loop=self.loop)
- delete_timeout -= 5
- else: # delete_timeout <= 0:
- raise ROclient.ROClientException(
- "Timeout waiting ns deleted from VIM"
- )
-
- except Exception as e:
- self.update_db_2("nsrs", nsr_id, db_nsr_update)
- if (
- isinstance(e, ROclient.ROClientException) and e.http_code == 404
- ): # not found
- db_nsr_update["_admin.deployed.RO.nsr_id"] = None
- db_nsr_update["_admin.deployed.RO.nsr_status"] = "DELETED"
- db_nsr_update["_admin.deployed.RO.nsr_delete_action_id"] = None
- self.logger.debug(
- logging_text + "RO_ns_id={} already deleted".format(ro_nsr_id)
- )
- elif (
- isinstance(e, ROclient.ROClientException) and e.http_code == 409
- ): # conflict
- failed_detail.append("delete conflict: {}".format(e))
- self.logger.debug(
- logging_text
- + "RO_ns_id={} delete conflict: {}".format(ro_nsr_id, e)
- )
- else:
- failed_detail.append("delete error: {}".format(e))
- self.logger.error(
- logging_text + "RO_ns_id={} delete error: {}".format(ro_nsr_id, e)
- )
-
- # Delete nsd
- if not failed_detail and deep_get(nsr_deployed, ("RO", "nsd_id")):
- ro_nsd_id = nsr_deployed["RO"]["nsd_id"]
- try:
- stage[2] = "Deleting nsd from RO."
- 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 self.RO.delete("nsd", ro_nsd_id)
- self.logger.debug(
- logging_text + "ro_nsd_id={} deleted".format(ro_nsd_id)
- )
- db_nsr_update["_admin.deployed.RO.nsd_id"] = None
- except Exception as e:
- if (
- isinstance(e, ROclient.ROClientException) and e.http_code == 404
- ): # not found
- db_nsr_update["_admin.deployed.RO.nsd_id"] = None
- self.logger.debug(
- logging_text + "ro_nsd_id={} already deleted".format(ro_nsd_id)
- )
- elif (
- isinstance(e, ROclient.ROClientException) and e.http_code == 409
- ): # conflict
- failed_detail.append(
- "ro_nsd_id={} delete conflict: {}".format(ro_nsd_id, e)
- )
- self.logger.debug(logging_text + failed_detail[-1])
- else:
- failed_detail.append(
- "ro_nsd_id={} delete error: {}".format(ro_nsd_id, e)
- )
- self.logger.error(logging_text + failed_detail[-1])
-
- if not failed_detail and deep_get(nsr_deployed, ("RO", "vnfd")):
- for index, vnf_deployed in enumerate(nsr_deployed["RO"]["vnfd"]):
- if not vnf_deployed or not vnf_deployed["id"]:
- continue
- try:
- ro_vnfd_id = vnf_deployed["id"]
- stage[
- 2
- ] = "Deleting member_vnf_index={} ro_vnfd_id={} from RO.".format(
- vnf_deployed["member-vnf-index"], ro_vnfd_id
- )
- 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 self.RO.delete("vnfd", ro_vnfd_id)
- self.logger.debug(
- logging_text + "ro_vnfd_id={} deleted".format(ro_vnfd_id)
- )
- db_nsr_update["_admin.deployed.RO.vnfd.{}.id".format(index)] = None
- except Exception as e:
- if (
- isinstance(e, ROclient.ROClientException) and e.http_code == 404
- ): # not found
- db_nsr_update[
- "_admin.deployed.RO.vnfd.{}.id".format(index)
- ] = None
- self.logger.debug(
- logging_text
- + "ro_vnfd_id={} already deleted ".format(ro_vnfd_id)
- )
- elif (
- isinstance(e, ROclient.ROClientException) and e.http_code == 409
- ): # conflict
- failed_detail.append(
- "ro_vnfd_id={} delete conflict: {}".format(ro_vnfd_id, e)
- )
- self.logger.debug(logging_text + failed_detail[-1])
- else:
- failed_detail.append(
- "ro_vnfd_id={} delete error: {}".format(ro_vnfd_id, e)
- )
- self.logger.error(logging_text + failed_detail[-1])
-
- if failed_detail:
- stage[2] = "Error deleting from VIM"
- else:
- stage[2] = "Deleted from VIM"
- db_nsr_update["detailed-status"] = " ".join(stage)
- self.update_db_2("nsrs", nsr_id, db_nsr_update)
- self._write_op_status(nslcmop_id, stage)
-
- if failed_detail:
- raise LcmException("; ".join(failed_detail))
-
async def terminate(self, nsr_id, nslcmop_id):
# Try to lock HA task here
task_is_locked_by_me = self.lcm_tasks.lock_HA("ns", "nslcmops", nslcmop_id)
) and vca.get("needed_terminate")
# For helm we must destroy_ee. Also for native_charm, as juju_model cannot be deleted if there are
# pending native charms
- destroy_ee = (
- True if vca_type in ("helm", "helm-v3", "native_charm") else False
- )
+ destroy_ee = True if vca_type in ("helm-v3", "native_charm") else False
# self.logger.debug(logging_text + "vca_index: {}, ee_id: {}, vca_type: {} destroy_ee: {}".format(
# vca_index, vca.get("ee_id"), vca_type, destroy_ee))
task = asyncio.ensure_future(
# 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."
logging_text, nsr_deployed, nsr_id, nslcmop_id, stage
)
)
- else:
- task_delete_ro = asyncio.ensure_future(
- self._terminate_RO(
- logging_text, nsr_deployed, nsr_id, nslcmop_id, stage
- )
- )
- tasks_dict_info[task_delete_ro] = "Removing deployment from VIM"
+ tasks_dict_info[task_delete_ro] = "Removing deployment from VIM"
# rest of staff will be done at finally
stage[1] = stage[2] = ""
except asyncio.CancelledError:
error_list.append("Cancelled")
- # TODO cancell all tasks
+ await self._cancel_pending_tasks(logging_text, tasks_dict_info)
+ await self._wait_for_tasks(
+ logging_text,
+ tasks_dict_info,
+ timeout_ns_terminate,
+ stage,
+ nslcmop_id,
+ )
except Exception as exc:
error_list.append(str(exc))
# update status at database
"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")
self._write_op_status(nslcmop_id, stage)
return error_detail_list
+ async def _cancel_pending_tasks(self, logging_text, created_tasks_info):
+ for task, name in created_tasks_info.items():
+ self.logger.debug(logging_text + "Cancelling task: " + name)
+ task.cancel()
+
@staticmethod
def _map_primitive_params(primitive_desc, params, instantiation_params):
"""
)
)
# wait and retry
- await asyncio.sleep(retries_interval, loop=self.loop)
+ await asyncio.sleep(retries_interval)
else:
if isinstance(e, asyncio.TimeoutError):
e = N2VCException(
nslcmop_operation_state = None
error_description_nslcmop = None
exc = None
+ step = ""
try:
# wait for any previous tasks in process
step = "Waiting for previous operations to terminate"
kdu_action = (
True
if primitive_name in actions
- and kdu["k8scluster-type"] not in ("helm-chart", "helm-chart-v3")
+ and kdu["k8scluster-type"] != "helm-chart-v3"
else False
)
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"
"nslcmop_id": nslcmop_id,
"operationState": nslcmop_operation_state,
},
- loop=self.loop,
)
except Exception as e:
self.logger.error(
raise
except Exception as e:
-
self.logger.debug("Error upgrading charm {}".format(path))
return "FAILED", "Error upgrading charm {}: {}".format(path, e)
exc = None
change_type = "updated"
detailed_status = ""
+ member_vnf_index = None
try:
# wait for any previous tasks in process
nsr_deployed = db_nsr["_admin"].get("deployed")
if update_type == "CHANGE_VNFPKG":
-
# Get the input parameters given through update request
vnf_instance_id = db_nslcmop["operationParams"][
"changeVnfPackageData"
step = "Checking if revision has changed in VNFD"
if current_vnf_revision != latest_vnfd_revision:
-
change_type = "policy_updated"
# There is new revision of VNFD, update operation is required
step = "Getting descriptor config"
if current_vnfd.get("kdu"):
-
search_key = "kdu_name"
else:
search_key = "vnfd_id"
# There could be several charm used in the same VNF
for ee_item in ee_list:
if ee_item.get("juju"):
-
step = "Getting charm name"
charm_name = ee_item["juju"].get("charm")
# add chart to list and all parameters
step = "Getting helm chart name"
chart_name = ee_item.get("helm-chart")
- if (
- ee_item.get("helm-version")
- and ee_item.get("helm-version") == "v2"
- ):
- vca_type = "helm"
- else:
- vca_type = "helm-v3"
+ vca_type = "helm-v3"
step = "Setting Helm chart artifact paths"
helm_artifacts.append(
if find_software_version(current_vnfd) != find_software_version(
latest_vnfd
):
-
step = "Checking if existing VNF has charm"
for current_charm_path, target_charm_path in list(
charm_artifact_paths
current_charm_path, target_charm_path
)
):
-
step = "Checking whether VNF uses juju bundle"
if check_juju_bundle_existence(current_vnfd):
-
raise LcmException(
"Charm upgrade is not supported for the instance which"
" uses juju-bundle: {}".format(
"nslcmop_id": nslcmop_id,
"operationState": nslcmop_operation_state,
}
- if change_type in ("vnf_terminated", "policy_updated"):
+ if (
+ change_type in ("vnf_terminated", "policy_updated")
+ 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)
if kdur.get("helm-chart"):
k8s_cluster_type = "helm-chart-v3"
self.logger.debug("kdur: {}".format(kdur))
- if (
- kdur.get("helm-version")
- and kdur.get("helm-version") == "v2"
- ):
- k8s_cluster_type = "helm-chart"
elif kdur.get("juju-bundle"):
k8s_cluster_type = "juju-bundle"
else:
if kdur.get("helm-chart"):
k8s_cluster_type = "helm-chart-v3"
self.logger.debug("kdur: {}".format(kdur))
- if (
- kdur.get("helm-version")
- and kdur.get("helm-version") == "v2"
- ):
- k8s_cluster_type = "helm-chart"
elif kdur.get("juju-bundle"):
k8s_cluster_type = "juju-bundle"
else:
exc_info=True,
)
finally:
+ error_list = list()
+ if exc:
+ error_list.append(str(exc))
self._write_ns_status(
nsr_id=nsr_id,
ns_state=None,
current_operation="IDLE",
current_operation_id=None,
)
- if tasks_dict_info:
- stage[1] = "Waiting for instantiate pending tasks."
- self.logger.debug(logging_text + stage[1])
- exc = await self._wait_for_tasks(
+ try:
+ if tasks_dict_info:
+ stage[1] = "Waiting for instantiate pending tasks."
+ self.logger.debug(logging_text + stage[1])
+ exc = await self._wait_for_tasks(
+ logging_text,
+ tasks_dict_info,
+ self.timeout.ns_deploy,
+ stage,
+ nslcmop_id,
+ nsr_id=nsr_id,
+ )
+ except asyncio.CancelledError:
+ error_list.append("Cancelled")
+ await self._cancel_pending_tasks(logging_text, tasks_dict_info)
+ await self._wait_for_tasks(
logging_text,
tasks_dict_info,
self.timeout.ns_deploy,
nslcmop_id,
nsr_id=nsr_id,
)
- if exc:
+ if error_list:
+ error_detail = "; ".join(error_list)
db_nslcmop_update[
"detailed-status"
- ] = error_description_nslcmop = "FAILED {}: {}".format(step, exc)
+ ] = error_description_nslcmop = "FAILED {}: {}".format(
+ step, error_detail
+ )
nslcmop_operation_state = "FAILED"
if db_nsr:
db_nsr_update["operational-status"] = old_operational_status
db_nsr_update[
"detailed-status"
] = "FAILED scaling nslcmop={} {}: {}".format(
- nslcmop_id, step, exc
+ nslcmop_id, step, error_detail
)
else:
error_description_nslcmop = None
"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)
vnfr_id: str,
nsr_id: str,
target_ip: str,
+ element_type: str,
vnf_member_index: str = "",
vdu_id: str = "",
vdu_index: int = None,
vnfr_id (str): VNFR ID where this EE applies
nsr_id (str): NSR ID where this EE applies
target_ip (str): VDU/KDU instance IP address
+ element_type (str): NS or VNF or VDU or KDU
vnf_member_index (str, optional): VNF index where this EE applies. Defaults to "".
vdu_id (str, optional): VDU ID where this EE applies. Defaults to "".
vdu_index (int, optional): VDU index where this EE applies. Defaults to None.
Returns:
_type_: Prometheus jobs
"""
- self.logger.debug(f"KDU: {kdu_name}; KDU INDEX: {kdu_index}")
+ # default the vdur and kdur names to an empty string, to avoid any later
+ # problem with Prometheus when the element type is not VDU or KDU
+ vdur_name = ""
+ kdur_name = ""
+
# look if exist a file called 'prometheus*.j2' and
artifact_content = self.fs.dir_ls(artifact_path)
job_file = next(
)
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()
- vdur_name = ""
- kdur_name = ""
- for r in range(360):
- db_vnfr = self.db.get_one("vnfrs", {"_id": vnfr_id})
- if vdu_id and vdu_index is not None:
- vdur = next(
- (
- x
- for x in get_iterable(db_vnfr, "vdur")
- if (
- x.get("vdu-id-ref") == vdu_id
- and x.get("count-index") == vdu_index
- )
- ),
- {},
- )
- if vdur.get("name"):
- vdur_name = vdur.get("name")
- break
- if kdu_name and kdu_index is not None:
- kdur = next(
- (
- x
- for x in get_iterable(db_vnfr, "kdur")
- if (
- x.get("kdu-name") == kdu_name
- and x.get("count-index") == kdu_index
- )
- ),
- {},
- )
- if kdur.get("name"):
- kdur_name = kdur.get("name")
- break
+ # obtain the VDUR or KDUR, if the element type is VDU or KDU
+ if element_type in ("VDU", "KDU"):
+ for _ in range(360):
+ db_vnfr = self.db.get_one("vnfrs", {"_id": vnfr_id})
+ if vdu_id and vdu_index is not None:
+ vdur = next(
+ (
+ x
+ for x in get_iterable(db_vnfr, "vdur")
+ if (
+ x.get("vdu-id-ref") == vdu_id
+ and x.get("count-index") == vdu_index
+ )
+ ),
+ {},
+ )
+ if vdur.get("name"):
+ vdur_name = vdur.get("name")
+ break
+ if kdu_name and kdu_index is not None:
+ kdur = next(
+ (
+ x
+ for x in get_iterable(db_vnfr, "kdur")
+ if (
+ x.get("kdu-name") == kdu_name
+ and x.get("count-index") == kdu_index
+ )
+ ),
+ {},
+ )
+ if kdur.get("name"):
+ kdur_name = kdur.get("name")
+ break
+
+ await asyncio.sleep(10)
+ else:
+ if vdu_id and vdu_index is not None:
+ raise LcmException(
+ f"Timeout waiting VDU with name={vdu_id} and index={vdu_index} to be intantiated"
+ )
+ if kdu_name and kdu_index is not None:
+ raise LcmException(
+ f"Timeout waiting KDU with name={kdu_name} and index={kdu_index} to be intantiated"
+ )
- await asyncio.sleep(10, loop=self.loop)
+ 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:
- if vdu_id and vdu_index is not None:
- raise LcmException(
- f"Timeout waiting VDU with name={vdu_id} and index={vdu_index} to be intantiated"
- )
- if kdu_name and kdu_index is not None:
- raise LcmException(
- 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,
- }
+ 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:
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
"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)
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")
exc_info=True,
)
finally:
- if tasks_dict_info:
- stage[1] = "Waiting for healing pending tasks."
- self.logger.debug(logging_text + stage[1])
- exc = await self._wait_for_tasks(
+ error_list = list()
+ if exc:
+ error_list.append(str(exc))
+ try:
+ if tasks_dict_info:
+ stage[1] = "Waiting for healing pending tasks."
+ self.logger.debug(logging_text + stage[1])
+ exc = await self._wait_for_tasks(
+ logging_text,
+ tasks_dict_info,
+ self.timeout.ns_deploy,
+ stage,
+ nslcmop_id,
+ nsr_id=nsr_id,
+ )
+ except asyncio.CancelledError:
+ error_list.append("Cancelled")
+ await self._cancel_pending_tasks(logging_text, tasks_dict_info)
+ await self._wait_for_tasks(
logging_text,
tasks_dict_info,
self.timeout.ns_deploy,
nslcmop_id,
nsr_id=nsr_id,
)
- if exc:
+ if error_list:
+ error_detail = "; ".join(error_list)
db_nslcmop_update[
"detailed-status"
- ] = error_description_nslcmop = "FAILED {}: {}".format(step, exc)
+ ] = error_description_nslcmop = "FAILED {}: {}".format(
+ step, error_detail
+ )
nslcmop_operation_state = "FAILED"
if db_nsr:
db_nsr_update["operational-status"] = old_operational_status
db_nsr_update["config-status"] = old_config_status
db_nsr_update[
"detailed-status"
- ] = "FAILED healing nslcmop={} {}: {}".format(nslcmop_id, step, exc)
+ ] = "FAILED healing nslcmop={} {}: {}".format(
+ nslcmop_id, step, error_detail
+ )
for task, task_name in tasks_dict_info.items():
if not task.done() or task.cancelled() or task.exception():
if task_name.startswith(self.task_name_deploy_vca):
"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)
vca_type = "native_charm"
elif ee_item.get("helm-chart"):
vca_name = ee_item["helm-chart"]
- if ee_item.get("helm-version") and ee_item.get("helm-version") == "v2":
- vca_type = "helm"
- else:
- vca_type = "helm-v3"
+ vca_type = "helm-v3"
else:
self.logger.debug(
logging_text + "skipping non juju neither charm configuration"
}
step = ""
try:
-
element_type = "NS"
element_under_configuration = nsr_id
# if SSH access is required, then get execution environment SSH public
# if native charm we have waited already to VM be UP
- if vca_type in ("k8s_proxy_charm", "lxc_proxy_charm", "helm", "helm-v3"):
+ if vca_type in ("k8s_proxy_charm", "lxc_proxy_charm", "helm-v3"):
pub_key = None
user = None
# self.logger.debug("get ssh key block")
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")
"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)
projects / osm / LCM.git / blobdiff