def update_config(tf_config: TerraformConfig,
cluster_config: ClusterConfig,
operators=None):
if tf_config is None:
tf_config = TerraformConfig()
if cluster_config is None:
cluster_config = ClusterConfig()
if operators is None:
operators = parse_olm_operators_from_env()
tf_config.worker_memory = resource_param(
tf_config.worker_memory, OperatorResource.WORKER_MEMORY_KEY,
operators)
tf_config.master_memory = resource_param(
tf_config.master_memory, OperatorResource.MASTER_MEMORY_KEY,
operators)
tf_config.worker_vcpu = resource_param(
tf_config.worker_vcpu, OperatorResource.WORKER_VCPU_KEY,
operators)
tf_config.master_vcpu = resource_param(
tf_config.master_vcpu, OperatorResource.MASTER_VCPU_KEY,
operators)
tf_config.workers_count = resource_param(
tf_config.workers_count, OperatorResource.WORKER_COUNT_KEY,
operators)
tf_config.worker_disk = resource_param(
tf_config.worker_disk, OperatorResource.WORKER_DISK_KEY,
operators)
tf_config.master_disk = resource_param(
tf_config.master_disk, OperatorResource.MASTER_DISK_KEY,
operators)
tf_config.master_disk_count = resource_param(
tf_config.master_disk_count,
OperatorResource.MASTER_DISK_COUNT_KEY, operators)
tf_config.worker_disk_count = resource_param(
tf_config.worker_disk_count,
OperatorResource.WORKER_DISK_COUNT_KEY, operators)
tf_config.nodes_count = tf_config.masters_count + tf_config.workers_count
cluster_config.olm_operators = [operators]
def start_install_and_wait_for_installed(self):
cluster_name = self.config.day1_cluster_name
# Running twice as a workaround for an issue with terraform not spawning a new node on first apply.
for _ in range(2):
with utils.file_lock_context():
utils.run_command(
f"make _apply_terraform CLUSTER_NAME={cluster_name} PLATFORM={consts.Platforms.BARE_METAL}"
)
time.sleep(5)
num_nodes_to_wait = self.config.day2_workers_count
installed_status = consts.NodesStatus.DAY2_INSTALLED
tfvars = utils.get_tfvars(self.config.tf_folder)
tf_network_name = tfvars["libvirt_network_name"]
config = TerraformConfig()
config.nodes_count = num_nodes_to_wait
libvirt_controller = LibvirtController(config=config,
entity_config=ClusterConfig())
libvirt_controller.wait_till_nodes_are_ready(
network_name=tf_network_name)
# Wait for day2 nodes
waiting.wait(
lambda: self.are_libvirt_nodes_in_cluster_hosts(),
timeout_seconds=consts.NODES_REGISTERED_TIMEOUT,
sleep_seconds=10,
waiting_for="Nodes to be registered in inventory service",
)
self.set_nodes_hostnames_if_needed(tf_network_name)
wait_till_all_hosts_are_in_status(
client=self.api_client,
cluster_id=self.config.cluster_id,
nodes_count=self.config.day2_workers_count,
statuses=[consts.NodesStatus.KNOWN],
interval=30,
)
# Start day2 nodes installation
log.info("Start installing all known nodes in the cluster %s",
self.config.cluster_id)
kubeconfig = utils.get_kubeconfig_path(self.config.day1_cluster_name)
ocp_ready_nodes = self.get_ocp_cluster_ready_nodes_num(kubeconfig)
hosts = self.api_client.get_cluster_hosts(self.config.cluster_id)
[
self.api_client.install_day2_host(self.config.infra_env_id,
host["id"]) for host in hosts
if host["status"] == "known"
]
log.info(
"Waiting until all nodes of cluster %s have been installed (reached added-to-existing-cluster)",
self.config.cluster_id,
)
wait_till_all_hosts_are_in_status(
client=self.api_client,
cluster_id=self.config.cluster_id,
nodes_count=num_nodes_to_wait,
statuses=[installed_status],
interval=30,
)
log.info(
"Waiting until installed nodes has actually been added to the OCP cluster"
)
waiting.wait(
lambda: self.wait_nodes_join_ocp_cluster(
ocp_ready_nodes, self.config.day2_workers_count, kubeconfig),
timeout_seconds=consts.NODES_REGISTERED_TIMEOUT,
sleep_seconds=30,
waiting_for="Day2 nodes to be added to OCP cluster",
expected_exceptions=Exception,
)
log.info("%d worker nodes were successfully added to OCP cluster",
self.config.day2_workers_count)