From e19bd74725c8bc7dabe721b1257279a8d1fa89ea Mon Sep 17 00:00:00 2001 From: tierno Date: Thu, 5 Dec 2019 16:09:08 +0000 Subject: [PATCH] workaround for bug 951. Allow k8scluster with only helm or juju Change-Id: I117a1415a1c1ce21736ff2dd0b0b3a9db81a83f9 Signed-off-by: tierno --- osm_lcm/vim_sdn.py | 117 ++++++++++++++++++++++++--------------------- 1 file changed, 63 insertions(+), 54 deletions(-) diff --git a/osm_lcm/vim_sdn.py b/osm_lcm/vim_sdn.py index 27ed850..931011b 100644 --- a/osm_lcm/vim_sdn.py +++ b/osm_lcm/vim_sdn.py @@ -24,6 +24,7 @@ from osm_lcm import ROclient from osm_lcm.lcm_utils import LcmException, LcmBase, deep_get from n2vc.k8s_helm_conn import K8sHelmConnector from n2vc.k8s_juju_conn import K8sJujuConnector +from n2vc.exceptions import K8sException, N2VCException from osm_common.dbbase import DbException from copy import deepcopy @@ -971,7 +972,6 @@ class K8sClusterLcm(LcmBase): db_k8scluster = None db_k8scluster_update = {} - db_juju_k8scluster_update = {} exc = None operationState_HA = '' @@ -982,46 +982,64 @@ class K8sClusterLcm(LcmBase): db_k8scluster = self.db.get_one("k8sclusters", {"_id": k8scluster_id}) self.db.encrypt_decrypt_fields(db_k8scluster.get("credentials"), 'decrypt', ['password', 'secret'], schema_version=db_k8scluster["schema_version"], salt=db_k8scluster["_id"]) - - k8s_hc_id, uninstall_sw = await self.helm_k8scluster.init_env( - yaml.safe_dump(db_k8scluster.get("credentials")) - ) - db_k8scluster_update["_admin.helm-chart.id"] = k8s_hc_id - db_k8scluster_update["_admin.helm-chart.created"] = uninstall_sw + k8s_credentials = yaml.safe_dump(db_k8scluster.get("credentials")) + error_text_list = [] + # helm-chart + k8s_hc_id = None + try: + k8s_hc_id, uninstall_sw = await self.helm_k8scluster.init_env(k8s_credentials) + db_k8scluster_update["_admin.helm-chart.id"] = k8s_hc_id + db_k8scluster_update["_admin.helm-chart.created"] = uninstall_sw + except Exception as e: + error_text_list.append("Failing init helm-chart: {}".format(e)) + db_k8scluster_update["_admin.helm-chart.error_msg"] = str(e) + if isinstance(e, K8sException): + self.logger.error(logging_text + "Failing init helm-chart: {}".format(e)) + else: + self.logger.error(logging_text + "Failing init helm-chart: {}".format(e), exc_info=True) # Juju/k8s cluster - k8s_jb_id, uninstall_sw = await self.juju_k8scluster.init_env( - yaml.safe_dump(db_k8scluster.get("credentials")) - ) - db_k8scluster_update["_admin.juju-bundle.id"] = k8s_jb_id - db_k8scluster_update["_admin.juju-bundle.created"] = uninstall_sw + k8s_jb_id = None + try: + k8s_jb_id, uninstall_sw = await self.juju_k8scluster.init_env(k8s_credentials) + db_k8scluster_update["_admin.juju-bundle.id"] = k8s_jb_id + db_k8scluster_update["_admin.juju-bundle.created"] = uninstall_sw + except Exception as e: + error_text_list.append("Failing init juju-bundle: {}".format(e)) + db_k8scluster_update["_admin.juju-bundle.error_msg"] = str(e) + if isinstance(e, N2VCException): + self.logger.error(logging_text + "Failing init juju-bundle: {}".format(e)) + else: + self.logger.error(logging_text + "Failing init juju-bundle: {}".format(e), exc_info=True) step = "Getting the list of repos" - self.logger.debug(logging_text + step) - task_list = [] - db_k8srepo_list = self.db.get_list("k8srepos", {}) - for repo in db_k8srepo_list: - step = "Adding repo {} to cluster: {}".format(repo["name"], k8s_hc_id) + if k8s_hc_id: self.logger.debug(logging_text + step) - task = asyncio.ensure_future(self.helm_k8scluster.repo_add(cluster_uuid=k8s_hc_id, - name=repo["name"], url=repo["url"], - repo_type="chart")) - task_list.append(task) - if not repo["_admin"].get("cluster-inserted"): - repo["_admin"]["cluster-inserted"] = [] - repo["_admin"]["cluster-inserted"].append(k8s_hc_id) - self.update_db_2("k8srepos", repo["_id"], repo) - - done = None - pending = None - if len(task_list) > 0: - self.logger.debug('Waiting for terminate pending tasks...') - done, pending = await asyncio.wait(task_list, timeout=3600) - if not pending: - self.logger.debug('All tasks finished...') - else: - self.logger.info('There are pending tasks: {}'.format(pending)) - db_k8scluster_update["_admin.operationalState"] = "ENABLED" + task_list = [] + db_k8srepo_list = self.db.get_list("k8srepos", {"type": "helm-chart"}) + for repo in db_k8srepo_list: + step = "Adding repo {} to cluster: {}".format(repo["name"], k8s_hc_id) + self.logger.debug(logging_text + step) + task = asyncio.ensure_future(self.helm_k8scluster.repo_add(cluster_uuid=k8s_hc_id, + name=repo["name"], url=repo["url"], + repo_type="chart")) + task_list.append(task) + repo_k8scluster_list = deep_get(repo, ("_admin", "cluster-inserted")) or [] + repo_k8scluster_list.append(k8s_hc_id) + self.update_db_2("k8srepos", repo["_id"], {"_admin.cluster-inserted": repo_k8scluster_list}) + + if task_list: + self.logger.debug(logging_text + 'Waiting for terminate tasks of repo_add') + done, pending = await asyncio.wait(task_list, timeout=3600) + if pending: + self.logger.error(logging_text + 'There are pending tasks: {}'.format(pending)) + + # mark as an error if both helm-chart and juju-bundle have been failed + if k8s_hc_id or k8s_jb_id: + db_k8scluster_update["_admin.operationalState"] = "ENABLED" + else: + db_k8scluster_update["_admin.operationalState"] = "ERROR" + db_k8scluster_update["_admin.detailed-status"] = ";".join(error_text_list) except Exception as e: self.logger.critical(logging_text + "Exit Exception {}".format(e), exc_info=True) @@ -1031,10 +1049,6 @@ class K8sClusterLcm(LcmBase): db_k8scluster_update["_admin.operationalState"] = "ERROR" db_k8scluster_update["_admin.detailed-status"] = "ERROR {}: {}".format(step, exc) - if db_juju_k8scluster_update: - db_juju_k8scluster_update["_admin.operationalState"] = "ERROR" - db_juju_k8scluster_update["_admin.detailed-status"] = "ERROR {}: {}".format(step, exc) - # Mark the k8scluster 'create' HA task as erroneous operationState_HA = 'FAILED' detailed_status_HA = "ERROR {}: {}".format(step, exc) @@ -1042,9 +1056,6 @@ class K8sClusterLcm(LcmBase): if db_k8scluster_update: self.update_db_2("k8sclusters", k8scluster_id, db_k8scluster_update) - if db_juju_k8scluster_update: - self.update_db_2("k8sclusters", k8scluster_id, db_juju_k8scluster_update) - # Register the K8scluster 'create' HA task either # succesful or erroneous, or do nothing (if legacy NBI) self.lcm_tasks.register_HA('k8scluster', 'create', op_id, @@ -1091,19 +1102,17 @@ class K8sClusterLcm(LcmBase): uninstall_sw = uninstall_sw or False cluster_removed = await self.juju_k8scluster.reset(cluster_uuid=k8s_jb_id, uninstall_sw=uninstall_sw) - if cluster_removed: - step = "Removing k8scluster='{}' from db".format(k8scluster_id) + if k8s_hc_id and cluster_removed: + step = "Removing k8scluster='{}' from k8srepos".format(k8scluster_id) self.logger.debug(logging_text + step) - db_k8srepo_list = self.db.get_list("k8srepos", {}) + db_k8srepo_list = self.db.get_list("k8srepos", {"_admin.cluster-inserted": k8s_hc_id}) for k8srepo in db_k8srepo_list: - index = 0 - for cluster in k8srepo["_admin"]["cluster-inserted"]: - hc_id = deep_get(db_k8scluster, ("_admin", "helm-chart", "id")) - if hc_id == cluster: - del(k8srepo["_admin"]["cluster-inserted"][index]) - break - index += 1 - self.update_db_2("k8srepos", k8srepo["_id"], k8srepo) + try: + cluster_list = k8srepo["_admin"]["cluster-inserted"] + cluster_list.remove(k8s_hc_id) + self.update_db_2("k8srepos", k8srepo["_id"], {"_admin.cluster-inserted": cluster_list}) + except Exception as e: + self.logger.error("{}: {}".format(step, e)) self.db.del_one("k8sclusters", {"_id": k8scluster_id}) else: raise LcmException("An error happened during the reset of the k8s cluster '{}'".format(k8scluster_id)) -- 2.17.1