updated default creation and gitignore. Fixed non-vital bug that didn…

…'t reset hosts for new cluster start.

.gitignore

@@ -4,6 +4,7 @@
 
 # variable resources
 resources/playbook/ansible.cfg
+resources/playbook/roles/bibigrid/templates/slurm/slurm.conf
 resources/playbook/site.yml
 resources/playbook/ansible_hosts
 resources/playbook/vars/

bibigrid/core/actions/create.py

@@ -59,6 +59,13 @@ def get_identifier(identifier, cluster_id, additional=""):
 WIREGUARD_SECURITY_GROUP_NAME = "wireguard" + SEPARATOR + "{cluster_id}"
 
 
+def create_defaults():
+    if not os.path.isfile(a_rp.ANSIBLE_CFG_PATH):
+        shutil.copy(a_rp.ANSIBLE_CFG_DEFAULT_PATH, a_rp.ANSIBLE_CFG_PATH)
+    if not os.path.isfile(a_rp.SLURM_CONF_TEMPLATE_PATH):
+        shutil.copy(a_rp.SLURM_CONF_TEMPLATE_DEFAULT_PATH, a_rp.SLURM_CONF_TEMPLATE_PATH)
+
+
 class Create:  # pylint: disable=too-many-instance-attributes,too-many-arguments
     """
     The class Create holds necessary methods to execute the Create-Action
@@ -211,9 +218,9 @@ def start_vpn_or_master(self, configuration, provider):
             configuration["floating_ip"] = server["private_v4"]  # pylint: enable=comparison-with-callable
         configuration["volumes"] = provider.get_mount_info_from_server(server)
 
-    def start_worker(self, worker, worker_count, configuration, provider):
+    def start_workers(self, worker, worker_count, configuration, provider):
         name = WORKER_IDENTIFIER(cluster_id=self.cluster_id, additional=worker_count)
-        self.log.info(f"Starting instance/server {name} on {provider.cloud_specification['identifier']}")
+        self.log.info(f"Starting worker {name} on {provider.cloud_specification['identifier']}.")
         flavor = worker["type"]
         network = configuration["network"]
         image = image_selection.select_image(provider, worker["image"], self.log,
@@ -222,12 +229,13 @@ def start_worker(self, worker, worker_count, configuration, provider):
         # create a server and block until it is up and running
         server = provider.create_server(name=name, flavor=flavor, key_name=self.key_name, image=image, network=network,
                                         volumes=None, security_groups=configuration["security_groups"], wait=True)
+        self.log.info(f"Worker {name} started on {provider.cloud_specification['identifier']}.")
         with self.worker_thread_lock:
             self.permanents.append(name)
             with open(a_rp.HOSTS_FILE, mode="r", encoding="utf-8") as hosts_file:
                 hosts = yaml.safe_load(hosts_file)
             if not hosts or "host_entries" not in hosts:
-                self.log.info(f"Resetting host entries because {'first run' if hosts else 'broken'}.")
+                self.log.warning("Hosts file is broken.")
                 hosts = {"host_entries": {}}
             hosts["host_entries"][name] = server["private_v4"]
             ansible_configurator.write_yaml(a_rp.HOSTS_FILE, hosts, self.log)
@@ -357,6 +365,7 @@ def start_start_server_threads(self):
         """
         start_server_threads = []
         worker_count = 0
+        ansible_configurator.write_yaml(a_rp.HOSTS_FILE, {"host_entries": {}}, self.log)
         for configuration, provider in zip(self.configurations, self.providers):
             start_server_thread = return_threading.ReturnThread(target=self.start_vpn_or_master,
                                                                 args=[configuration, provider])
@@ -365,7 +374,7 @@ def start_start_server_threads(self):
             for worker in configuration.get("workerInstances", []):
                 if not worker.get("onDemand", True):
                     for _ in range(int(worker["count"])):
-                        start_server_thread = return_threading.ReturnThread(target=self.start_worker,
+                        start_server_thread = return_threading.ReturnThread(target=self.start_workers,
                                                                             args=[worker, worker_count, configuration,
                                                                                   provider])
                         start_server_thread.start()
@@ -414,8 +423,7 @@ def create(self):  # pylint: disable=too-many-branches,too-many-statements
         try:
             self.generate_keypair()
             self.prepare_configurations()
-            if not os.path.isfile(a_rp.ANSIBLE_CFG_PATH):
-                shutil.copy(a_rp.ANSIBLE_CFG_DEFAULT_PATH, a_rp.ANSIBLE_CFG_PATH)
+            create_defaults()
             self.generate_security_groups()
             self.start_start_server_threads()
             self.extended_network_configuration()

bibigrid/core/utility/paths/ansible_resources_path.py

@@ -23,6 +23,7 @@
 ADDITIONAL_ROLES_PATH: str = ROLES_PATH + "additional/"
 DEFAULT_IP_FILE = VARS_PATH + "{{ ansible_default_ipv4.address }}.yml"
 ANSIBLE_CFG = "ansible.cfg"
+SLURM_CONF = "slurm.conf"
 
 # LOCAL
 PLAYBOOK = "playbook/"
@@ -40,8 +41,12 @@
 VARS_FOLDER = os.path.join(PLAYBOOK_PATH, VARS_PATH)
 GROUP_VARS_FOLDER = os.path.join(PLAYBOOK_PATH, GROUP_VARS_PATH)
 HOST_VARS_FOLDER = os.path.join(PLAYBOOK_PATH, HOST_VARS_PATH)
-## DEFAULTS
-ANSIBLE_CFG_DEFAULT_PATH = os.path.join(b_p.RESOURCES_PATH, "defaults", "ansible", ANSIBLE_CFG)
+SLURM_CONF_TEMPLATE_PATH = os.path.join(PLAYBOOK_PATH, "roles", "bibigrid", "templates", "slurm", SLURM_CONF)
+
+# DEFAULTS
+DEFAULTS = os.path.join(b_p.RESOURCES_PATH, "defaults")
+ANSIBLE_CFG_DEFAULT_PATH = os.path.join(DEFAULTS, "ansible", ANSIBLE_CFG)
+SLURM_CONF_TEMPLATE_DEFAULT_PATH = os.path.join(DEFAULTS, "slurm", SLURM_CONF)
 
 
 # REMOTE