def status(args): # noqa: C901 FIXME!!!
stack_name = utils.get_stack_name(args.cluster_name)
# Parse configuration file to read the AWS section
PclusterConfig.init_aws(config_file=args.config_file)
cfn = boto3.client("cloudformation")
try:
stack = utils.get_stack(stack_name, cfn)
sys.stdout.write("\rStatus: %s" % stack.get("StackStatus"))
sys.stdout.flush()
if not args.nowait:
while stack.get("StackStatus") not in [
"CREATE_COMPLETE",
"UPDATE_COMPLETE",
"UPDATE_ROLLBACK_COMPLETE",
"ROLLBACK_COMPLETE",
"CREATE_FAILED",
"DELETE_FAILED",
]:
time.sleep(5)
stack = utils.get_stack(stack_name, cfn)
events = utils.get_stack_events(stack_name)[0]
resource_status = (
"Status: %s - %s" % (events.get("LogicalResourceId"), events.get("ResourceStatus"))
).ljust(80)
sys.stdout.write("\r%s" % resource_status)
sys.stdout.flush()
sys.stdout.write("\rStatus: %s\n" % stack.get("StackStatus"))
sys.stdout.flush()
if stack.get("StackStatus") in ["CREATE_COMPLETE", "UPDATE_COMPLETE", "UPDATE_ROLLBACK_COMPLETE"]:
state = _poll_head_node_state(stack_name)
if state == "running":
_print_stack_outputs(stack)
_print_compute_fleet_status(args.cluster_name, stack)
elif stack.get("StackStatus") in ["ROLLBACK_COMPLETE", "CREATE_FAILED", "DELETE_FAILED"]:
events = utils.get_stack_events(stack_name)
for event in events:
if event.get("ResourceStatus") in ["CREATE_FAILED", "DELETE_FAILED", "UPDATE_FAILED"]:
LOGGER.info(
"%s %s %s %s %s",
event.get("Timestamp"),
event.get("ResourceStatus"),
event.get("ResourceType"),
event.get("LogicalResourceId"),
event.get("ResourceStatusReason"),
)
else:
sys.stdout.write("\n")
sys.stdout.flush()
except ClientError as e:
LOGGER.critical(e.response.get("Error").get("Message"))
sys.stdout.flush()
sys.exit(1)
except KeyboardInterrupt:
LOGGER.info("\nExiting...")
sys.exit(0)
def delete(args):
saw_update = False
LOGGER.info("Deleting: %s", args.cluster_name)
stack_name = utils.get_stack_name(args.cluster_name)
# Parse configuration file to read the AWS section
PclusterConfig.init_aws(config_file=args.config_file)
cfn = boto3.client("cloudformation")
try:
# delete_stack does not raise an exception if stack does not exist
# Use describe_stacks to explicitly check if the stack exists
cfn.describe_stacks(StackName=stack_name)
cfn.delete_stack(StackName=stack_name)
saw_update = True
stack_status = utils.get_stack(stack_name, cfn).get("StackStatus")
sys.stdout.write("\rStatus: %s" % stack_status)
sys.stdout.flush()
LOGGER.debug("Status: %s", stack_status)
if not args.nowait:
while stack_status == "DELETE_IN_PROGRESS":
time.sleep(5)
stack_status = utils.get_stack(stack_name,
cfn).get("StackStatus")
events = cfn.describe_stack_events(
StackName=stack_name).get("StackEvents")[0]
resource_status = ("Status: %s - %s" %
(events.get("LogicalResourceId"),
events.get("ResourceStatus"))).ljust(80)
sys.stdout.write("\r%s" % resource_status)
sys.stdout.flush()
sys.stdout.write("\rStatus: %s\n" % stack_status)
sys.stdout.flush()
LOGGER.debug("Status: %s", stack_status)
else:
sys.stdout.write("\n")
sys.stdout.flush()
if stack_status == "DELETE_FAILED":
LOGGER.info(
"Cluster did not delete successfully. Run 'pcluster delete %s' again",
args.cluster_name)
except ClientError as e:
if e.response.get("Error").get("Message").endswith("does not exist"):
if saw_update:
LOGGER.info("\nCluster deleted successfully.")
sys.exit(0)
LOGGER.critical(e.response.get("Error").get("Message"))
sys.stdout.flush()
sys.exit(1)
except KeyboardInterrupt:
LOGGER.info("\nExiting...")
sys.exit(0)
def delete(args):
PclusterConfig.init_aws(config_file=args.config_file)
LOGGER.info("Deleting: %s", args.cluster_name)
stack_name = utils.get_stack_name(args.cluster_name)
if not utils.stack_exists(stack_name):
if args.keep_logs:
utils.warn(
"Stack for {0} does not exist. Cannot prevent its log groups from being deleted."
.format(args.cluster_name))
utils.warn("Cluster {0} has already been deleted.".format(
args.cluster_name))
sys.exit(0)
elif args.keep_logs:
_persist_cloudwatch_log_groups(args.cluster_name)
_delete_cluster(args.cluster_name, args.nowait)
def instances(args):
stack_name = utils.get_stack_name(args.cluster_name)
PclusterConfig.init_aws(config_file=args.config_file)
cfn_stack = utils.get_stack(stack_name)
scheduler = utils.get_cfn_param(cfn_stack.get("Parameters"), "Scheduler")
instances = []
head_node_server = utils.describe_cluster_instances(stack_name, node_type=utils.NodeType.head_node)
if head_node_server:
instances.append(("MasterServer", head_node_server[0].get("InstanceId")))
if scheduler != "awsbatch":
instances.extend(_get_compute_instances(stack_name))
for instance in instances:
LOGGER.info("%s %s", instance[0], instance[1])
if scheduler == "awsbatch":
LOGGER.info("Run 'awsbhosts --cluster %s' to list the compute instances", args.cluster_name)
def dcv_connect(args):
"""
Execute pcluster dcv connect command.
:param args: pcluster cli arguments.
"""
# Parse configuration file to read the AWS section
PclusterConfig.init_aws(
) # FIXME it always searches for the default configuration file
# Prepare ssh command to execute in the head node instance
stack = get_stack(get_stack_name(args.cluster_name))
shared_dir = get_cfn_param(stack.get("Parameters"), "SharedDir")
head_node_ip, username = get_head_node_ip_and_username(args.cluster_name)
cmd = 'ssh {CFN_USER}@{HEAD_NODE_IP} {KEY} "{REMOTE_COMMAND} {DCV_SHARED_DIR}"'.format(
CFN_USER=username,
HEAD_NODE_IP=head_node_ip,
KEY="-i {0}".format(args.key_path) if args.key_path else "",
REMOTE_COMMAND=DCV_CONNECT_SCRIPT,
DCV_SHARED_DIR=shared_dir,
)
try:
url = retry(_retrieve_dcv_session_url,
func_args=[cmd, args.cluster_name, head_node_ip],
attempts=4)
url_message = "Please use the following one-time URL in your browser within 30 seconds:\n{0}".format(
url)
except DCVConnectionError as e:
error("Something went wrong during DCV connection.\n{0}"
"Please check the logs in the /var/log/parallelcluster/ folder "
"of the head node and submit an issue {1}\n".format(
e, PCLUSTER_ISSUES_LINK))
if args.show_url:
LOGGER.info(url_message)
return
try:
if not webbrowser.open_new(url):
raise webbrowser.Error("Unable to open the Web browser.")
except webbrowser.Error as e:
LOGGER.info("{0}\n{1}".format(e, url_message))
def list_stacks(args):
# Parse configuration file to read the AWS section
PclusterConfig.init_aws(config_file=args.config_file)
try:
result = []
for stack in utils.paginate_boto3(boto3.client("cloudformation").describe_stacks):
if stack.get("ParentId") is None and stack.get("StackName").startswith(PCLUSTER_STACK_PREFIX):
pcluster_version = _get_pcluster_version_from_stack(stack)
result.append(
[
stack.get("StackName")[len(PCLUSTER_STACK_PREFIX) :], # noqa: E203
_colorize(stack.get("StackStatus"), args),
pcluster_version,
]
)
LOGGER.info(tabulate(result, tablefmt="plain"))
except ClientError as e:
LOGGER.critical(e.response.get("Error").get("Message"))
sys.exit(1)
except KeyboardInterrupt:
LOGGER.info("Exiting...")
sys.exit(0)
def dcv_connect(args):
"""
Execute pcluster dcv connect command.
:param args: pcluster cli arguments.
"""
# Parse configuration file to read the AWS section
PclusterConfig.init_aws(
) # FIXME it always searches for the default configuration file
# Prepare ssh command to execute in the master instance
stack = get_stack(get_stack_name(args.cluster_name))
shared_dir = get_cfn_param(stack.get("Parameters"), "SharedDir")
master_ip, username = get_master_ip_and_username(args.cluster_name)
cmd = 'ssh {CFN_USER}@{MASTER_IP} {KEY} "{REMOTE_COMMAND} {DCV_SHARED_DIR}"'.format(
CFN_USER=username,
MASTER_IP=master_ip,
KEY="-i {0}".format(args.key_path) if args.key_path else "",
REMOTE_COMMAND=DCV_CONNECT_SCRIPT,
DCV_SHARED_DIR=shared_dir,
)
# Connect by ssh to the master instance and prepare DCV session
try:
LOGGER.debug("SSH command: {0}".format(cmd))
output = _check_command_output(cmd)
# At first ssh connection, the ssh command alerts it is adding the host to the known hosts list
if re.search("Permanently added .* to the list of known hosts.",
output):
output = _check_command_output(cmd)
dcv_parameters = re.search(
r"PclusterDcvServerPort=([\d]+) PclusterDcvSessionId=([\w]+) PclusterDcvSessionToken=([\w-]+)",
output)
if dcv_parameters:
dcv_server_port = dcv_parameters.group(1)
dcv_session_id = dcv_parameters.group(2)
dcv_session_token = dcv_parameters.group(3)
else:
error(
"Something went wrong during DCV connection. Please manually execute the command:\n{0}\n"
"If the problem persists, please check the logs in the /var/log/parallelcluster/ folder "
"of the master instance and submit an issue {1}.".format(
cmd, PCLUSTER_ISSUES_LINK))
except sub.CalledProcessError as e:
if "{0}: No such file or directory".format(
DCV_CONNECT_SCRIPT) in e.output:
error(
"The cluster {0} has been created with an old version of ParallelCluster "
"without the DCV support.".format(args.cluster_name))
else:
error("Something went wrong during DCV connection.\n{0}".format(
e.output))
# Open web browser
url = "https://{IP}:{PORT}?authToken={TOKEN}#{SESSION_ID}".format(
IP=master_ip,
PORT=dcv_server_port,
TOKEN=dcv_session_token,
SESSION_ID=dcv_session_id)
try:
webbrowser.open_new(url)
except webbrowser.Error:
LOGGER.info(
"Unable to open the Web browser. "
"Please use the following URL in your browser within 30 seconds:\n{0}"
.format(url))
