def shutdown_via_ssh(node_list, ssh_file):
if ssh_file != "":
ssh_param = " -i " + ssh_file
else:
ssh_param = ""
for node in node_list:
logging.info("==== Shutdown " + str(node) + " ====")
run_cmd("ssh " + ssh_param + "core@" + node + " sudo shutdown - h")
def backup_etcd(master_node):
run_cmd(
"oc debug node/" + master_node +
" -- chroot /host /usr/local/bin/cluster-backup.sh /home/core/assets/backup"
)
def wait_for_ready_status(node, timeout):
run_cmd("kubectl wait --for=condition=Ready "
"node/" + node + " --timeout=" + str(timeout) + "s")
def approve_crs():
run_cmd("oc get csr | grep Pending | cut -f1 -d"
" | while read i; do oc adm certificate approve $i; done")
def main(cfg):
# Parse and read the config
if os.path.isfile(cfg):
with open(cfg, 'r') as f:
config = yaml.full_load(f)
config = config["shutdown"][0]
cloud_type = config['cloud_type']
kubeconfig_path = config.get("kubeconfig_path", "~/.kube/config")
shutdown_master_num = config.get("shutdown_master_num", "all")
shutdown_worker_num = config.get("shutdown_worker_num", "all")
shutdown_infra_num = config.get("shutdown_infra_num", "all")
ssh_file = config.get("ssh_file", "")
initialize_clients(kubeconfig_path)
downtime = calc_time(config.get("downtime", "300 s"))
masters = list_nodes("node-role.kubernetes.io/master")
backup_etcd(masters[1])
if shutdown_master_num != "all":
new_master_list = []
for i in range(int(shutdown_master_num)):
new_master_list.append(masters[i])
masters = new_master_list
workers = list_nodes(
"node-role.kubernetes.io/worker=,node-role.kubernetes.io/infra!=")
if shutdown_worker_num != "all":
new_worker_list = []
for i in range(int(shutdown_worker_num)):
new_worker_list.append(workers[i])
workers = new_worker_list
infras = list_nodes("node-role.kubernetes.io/infra")
if shutdown_infra_num != "all":
new_infra_list = []
for i in range(int(shutdown_infra_num)):
new_infra_list.append(infras[i])
infras = new_infra_list
node_list = workers + infras + masters
logging.info('node list ' + str(node_list))
if cloud_type == "aws":
aws = aws_node_scenarios()
for node in node_list:
logging.info('stop node ' + str(node))
aws_node_scenarios.node_stop_scenario(aws, node)
elif cloud_type == "azure" or cloud_type == "az":
logging.info("azure")
az_account = run_cmd("az account list -o yaml")
az = azure_node_scenarios(az_account)
for node in node_list:
logging.info('stop node ' + str(node))
azure_node_scenarios.node_stop_scenario(az, node)
elif cloud_type == "gcp":
logging.info('gcp')
project = run_cmd('gcloud config get-value project').split(
'/n')[0].strip()
gcp = gcp_node_scenarios(project)
for node in node_list:
logging.info('stop node ' + str(node))
gcp_node_scenarios.node_stop_scenario(gcp, node)
else:
logging.info("Shutting down using ssh")
shutdown_via_ssh(node_list, ssh_file)
# wait period
time.sleep(downtime)
# restart cluster
# start nodes based on cloud provider
if cloud_type == "aws":
for node in node_list:
logging.info('start node ' + str(node))
aws_node_scenarios.node_start_scenario(aws, node)
elif cloud_type == "azure" or cloud_type == "az":
for node in node_list:
logging.info('start node ' + str(node))
azure_node_scenarios.node_start_scenario(az, node)
elif cloud_type == "gcp":
for node in node_list:
logging.info('start node ' + str(node))
gcp_node_scenarios.node_start_scenario(gcp, node)
else:
logging.info("Cloud type " + str(cloud_type) +
" is not supported ")
sys.exit(1)
wait_for_all_nodes_ready(masters)
wait_for_all_nodes_ready(workers)
wait_for_all_nodes_ready(infras)
cluster_operators = run_cmd("oc get co")
run_cmd("oc get nodes")