def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster = spark_client.get_cluster(args.cluster_id)
cluster_config = spark_client.get_cluster_config(args.cluster_id)
ssh_conf = SshConfig()
ssh_conf.merge(
cluster_id=args.cluster_id,
username=args.username,
job_ui_port=args.jobui,
job_history_ui_port=args.jobhistoryui,
web_ui_port=args.webui,
host=args.host,
connect=args.connect,
internal=args.internal)
log.info("-------------------------------------------")
utils.log_property("spark cluster id", ssh_conf.cluster_id)
utils.log_property("open webui", "{0}{1}".format(http_prefix, ssh_conf.web_ui_port))
utils.log_property("open jobui", "{0}{1}".format(http_prefix, ssh_conf.job_ui_port))
utils.log_property("open jobhistoryui", "{0}{1}".format(http_prefix, ssh_conf.job_history_ui_port))
print_plugin_ports(cluster_config)
utils.log_property("ssh username", ssh_conf.username)
utils.log_property("connect", ssh_conf.connect)
log.info("-------------------------------------------")
try:
shell_out_ssh(spark_client, ssh_conf)
except OSError:
# no ssh client is found, falling back to pure python
native_python_ssh_into_master(spark_client, cluster, ssh_conf, args.password)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
log.info('-------------------------------------------')
log.info('spark cluster id: {}'.format(args.cluster_id))
log.info('username: {}'.format(args.username))
log.info('-------------------------------------------')
if args.ssh_key:
ssh_key = args.ssh_key
else:
ssh_key = spark_client.secrets_config.ssh_pub_key
ssh_key, password = utils.get_ssh_key_or_prompt(
ssh_key, args.username, args.password, spark_client.secrets_config)
spark_client.create_user(cluster_id=args.cluster_id,
username=args.username,
password=password,
ssh_key=ssh_key)
if password:
log.info('password: %s', '*' * len(password))
elif ssh_key:
log.info('ssh public key: %s', ssh_key)
log.info('-------------------------------------------')
def get_spark_client():
# load secrets
# note: this assumes secrets are set up in .aztk/secrets
tenant_id = os.environ.get("TENANT_ID")
client_id = os.environ.get("CLIENT_ID")
credential = os.environ.get("CREDENTIAL")
batch_account_resource_id = os.environ.get("BATCH_ACCOUNT_RESOURCE_ID")
storage_account_resource_id = os.environ.get("STORAGE_ACCOUNT_RESOURCE_ID")
ssh_pub_key = os.environ.get("ID_RSA_PUB")
ssh_private_key = os.environ.get("ID_RSA")
keys = [
tenant_id, client_id, credential, batch_account_resource_id,
storage_account_resource_id, ssh_private_key, ssh_pub_key
]
spark_client = None
if all(keys):
spark_client = aztk.spark.Client(
aztk.spark.models.SecretsConfiguration(
service_principal=aztk.spark.models.
ServicePrincipalConfiguration(
tenant_id=tenant_id,
client_id=client_id,
credential=credential,
batch_account_resource_id=batch_account_resource_id,
storage_account_resource_id=storage_account_resource_id),
ssh_pub_key=ssh_pub_key,
ssh_priv_key=ssh_private_key))
else:
# fallback to local secrets if environment variables don't exist
spark_client = aztk.spark.Client(config.load_aztk_secrets())
return spark_client
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
clusters = spark_client.list_clusters()
if args.quiet:
utils.print_clusters_quiet(clusters)
else:
utils.print_clusters(clusters)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster_ids = args.cluster_ids
for cluster_id in cluster_ids:
if not args.force:
if not args.keep_logs:
log.warning(
"All logs persisted for this cluster will be deleted.")
confirmation_cluster_id = input(
"Please confirm the id of the cluster you wish to delete [{}]: "
.format(cluster_id))
if confirmation_cluster_id != cluster_id:
log.error(
"Confirmation cluster id does not match. Please try again."
)
return
if spark_client.cluster.delete(id=cluster_id,
keep_logs=args.keep_logs):
log.info("Deleting cluster %s", cluster_id)
else:
log.error(
"Cluster with id '%s' doesn't exist or was already deleted.",
cluster_id)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
job_conf = JobConfig()
job_conf.merge(args.job_id, args.job_conf)
# by default, load spark configuration files in .aztk/
spark_configuration = config.load_aztk_spark_config()
# overwrite with values in job_conf if they exist
if job_conf.spark_defaults_conf:
spark_configuration.spark_defaults_conf = job_conf.spark_defaults_conf
if job_conf.spark_env_sh:
spark_configuration.spark_env_sh = job_conf.spark_env_sh
if job_conf.core_site_xml:
spark_configuration.core_site_xml = job_conf.core_site_xml
job_configuration = aztk.spark.models.JobConfiguration(
id=job_conf.id,
applications=job_conf.applications,
custom_scripts=job_conf.custom_scripts,
spark_configuration=spark_configuration,
vm_size=job_conf.vm_size,
docker_repo=job_conf.docker_repo,
max_dedicated_nodes=job_conf.max_dedicated_nodes,
max_low_pri_nodes=job_conf.max_low_pri_nodes,
subnet_id=job_conf.subnet_id,
worker_on_master=job_conf.worker_on_master)
#TODO: utils.print_job_conf(job_configuration)
spark_client.submit_job(job_configuration)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
job_conf = JobConfig()
job_conf.merge(args.job_id, args.job_conf)
# by default, load spark configuration files in .aztk/
spark_configuration = config.load_aztk_spark_config()
# overwrite with values in job_conf if they exist
if job_conf.spark_defaults_conf:
spark_configuration.spark_defaults_conf = job_conf.spark_defaults_conf
if job_conf.spark_env_sh:
spark_configuration.spark_env_sh = job_conf.spark_env_sh
if job_conf.core_site_xml:
spark_configuration.core_site_xml = job_conf.core_site_xml
job_configuration = aztk.spark.models.JobConfiguration(
id=job_conf.id,
applications=job_conf.applications,
spark_configuration=spark_configuration,
vm_size=job_conf.vm_size,
toolkit=job_conf.toolkit,
max_dedicated_nodes=job_conf.max_dedicated_nodes,
max_low_pri_nodes=job_conf.max_low_pri_nodes,
subnet_id=job_conf.subnet_id,
plugins=job_conf.plugins,
worker_on_master=job_conf.worker_on_master,
scheduling_target=job_conf.scheduling_target,
)
# TODO: utils.print_job_conf(job_configuration)
spark_client.job.submit(job_configuration, args.vm_os_ver)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
if spark_client.stop_job_app(args.job_id, args.app_name):
log.info("Stopped app {0}".format(args.app_name))
else:
log.error("App with name {0} does not exist or was already deleted")
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
log.info("-------------------------------------------")
log.info("spark cluster id: %s", args.cluster_id)
log.info("username: %s", args.username)
log.info("-------------------------------------------")
if args.ssh_key:
ssh_key = args.ssh_key
else:
ssh_key = spark_client.secrets_configuration.ssh_pub_key
ssh_key, password = utils.get_ssh_key_or_prompt(
ssh_key, args.username, args.password,
spark_client.secrets_configuration)
spark_client.cluster.create_user(id=args.cluster_id,
username=args.username,
password=password,
ssh_key=ssh_key)
if password:
log.info("password: %s", "*" * len(password))
elif ssh_key:
log.info("ssh public key: %s", ssh_key)
log.info("-------------------------------------------")
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
timestr = time.strftime("%Y%m%d-%H%M%S")
if not args.output:
args.output = os.path.join(os.getcwd(), "debug-{0}-{1}".format(args.cluster_id, timestr))
with utils.Spinner():
spark_client.cluster.diagnostics(id=args.cluster_id, output_directory=args.output, brief=args.brief)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
if spark_client.job.stop_application(args.job_id, args.app_name):
log.info("Stopped app %s", args.app_name)
else:
log.error("App with name %s does not exist or was already deleted",
args.app_name)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
with utils.Spinner():
copy_output = spark_client.cluster.copy(
id=args.cluster_id, source_path=args.source_path, destination_path=args.dest_path, internal=args.internal)
for node_output in copy_output:
utils.log_node_copy_output(node_output)
sys.exit(0 if not any([node_output.error for node_output in copy_output]) else 1)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster_conf = ClusterConfiguration()
cluster_conf.spark_configuration = load_aztk_spark_config(args.spark_conf)
# read cluster.yaml configuration file, overwrite values with args
file_config, wait = config.read_cluster_config() if args.cluster_path is None \
else config.read_cluster_config(args.cluster_path)
cluster_conf.merge(file_config)
cluster_conf.merge(
ClusterConfiguration(
cluster_id=args.cluster_id,
size=args.size,
size_low_priority=args.size_low_priority,
vm_size=args.vm_size,
subnet_id=args.subnet_id,
user_configuration=UserConfiguration(username=args.username,
password=args.password),
))
if cluster_conf.toolkit:
if args.docker_repo:
cluster_conf.toolkit.docker_repo = args.docker_repo
if args.docker_run_options:
cluster_conf.toolkit.docker_run_options = args.docker_run_options
wait = wait if args.wait is None else args.wait
user_configuration = cluster_conf.user_configuration
if user_configuration and user_configuration.username:
ssh_key, password = utils.get_ssh_key_or_prompt(
spark_client.secrets_configuration.ssh_pub_key,
user_configuration.username,
user_configuration.password,
spark_client.secrets_configuration,
)
cluster_conf.user_configuration = aztk.spark.models.UserConfiguration(
username=user_configuration.username,
password=password,
ssh_key=ssh_key)
else:
cluster_conf.user_configuration = None
cluster_conf.validate()
utils.print_cluster_conf(cluster_conf, wait)
with utils.Spinner():
# create spark cluster
cluster = spark_client.cluster.create(
cluster_configuration=cluster_conf,
vm_ver=args.vm_os_ver,
wait=wait)
if wait:
log.info("Cluster %s created successfully.", cluster.id)
else:
log.info("Cluster %s is being provisioned.", cluster.id)
def execute(args: typing.NamedTuple):
if not args.wait and args.output:
raise aztk.error.AztkError("--output flag requires --wait flag")
spark_client = aztk.spark.Client(config.load_aztk_secrets())
jars = []
py_files = []
files = []
if args.jars is not None:
jars = args.jars.replace(" ", "").split(",")
if args.py_files is not None:
py_files = args.py_files.replace(" ", "").split(",")
if args.files is not None:
files = args.files.replace(" ", "").split(",")
log_application(args, jars, py_files, files)
spark_client.cluster.submit(
id=args.cluster_id,
application=aztk.spark.models.ApplicationConfiguration(
name=args.name,
application=args.app,
application_args=args.app_args,
main_class=args.main_class,
jars=jars,
py_files=py_files,
files=files,
driver_java_options=args.driver_java_options,
driver_library_path=args.driver_library_path,
driver_class_path=args.driver_class_path,
driver_memory=args.driver_memory,
executor_memory=args.executor_memory,
driver_cores=args.driver_cores,
executor_cores=args.executor_cores,
max_retry_count=args.max_retry_count,
),
remote=args.remote,
internal=args.internal,
wait=False,
)
if args.wait:
if not args.output:
exit_code = utils.stream_logs(client=spark_client, cluster_id=args.cluster_id, application_name=args.name)
else:
with utils.Spinner():
spark_client.cluster.wait(id=args.cluster_id, application_name=args.name)
application_log = spark_client.cluster.get_application_log(
id=args.cluster_id, application_name=args.name)
with open(os.path.abspath(os.path.expanduser(args.output)), "w", encoding="UTF-8") as f:
f.write(application_log.log)
exit_code = application_log.exit_code
sys.exit(exit_code)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
with utils.Spinner():
if args.node_id:
results = [spark_client.node_run(args.cluster_id, args.node_id, args.command, args.host, args.internal)]
else:
results = spark_client.cluster_run(args.cluster_id, args.command, args.host, args.internal)
[utils.log_execute_result(node_id, result) for node_id, result in results]
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster = spark_client.get_cluster(args.cluster_id)
cluster_config = spark_client.get_cluster_config(args.cluster_id)
ssh_conf = SshConfig()
ssh_conf.merge(cluster_id=args.cluster_id,
username=args.username,
job_ui_port=args.jobui,
job_history_ui_port=args.jobhistoryui,
web_ui_port=args.webui,
jupyter_port=args.jupyter,
name_node_ui_port=args.namenodeui,
rstudio_server_port=args.rstudioserver,
host=args.host,
connect=args.connect)
log.info("-------------------------------------------")
utils.log_property("spark cluster id", ssh_conf.cluster_id)
utils.log_property("open webui", "{0}{1}".format(http_prefix,
ssh_conf.web_ui_port))
utils.log_property("open jobui", "{0}{1}".format(http_prefix,
ssh_conf.job_ui_port))
utils.log_property(
"open jobhistoryui", "{0}{1}".format(http_prefix,
ssh_conf.job_history_ui_port))
print_plugin_ports(cluster_config)
utils.log_property("ssh username", ssh_conf.username)
utils.log_property("connect", ssh_conf.connect)
log.info("-------------------------------------------")
# get ssh command
try:
ssh_cmd = utils.ssh_in_master(
client=spark_client,
cluster_id=ssh_conf.cluster_id,
webui=ssh_conf.web_ui_port,
jobui=ssh_conf.job_ui_port,
jobhistoryui=ssh_conf.job_history_ui_port,
username=ssh_conf.username,
host=ssh_conf.host,
connect=ssh_conf.connect)
if not ssh_conf.connect:
log.info("")
log.info(
"Use the following command to connect to your spark head node:"
)
log.info("\t%s", ssh_cmd)
except batch_error.BatchErrorException as e:
if e.error.code == "PoolNotFound":
raise aztk.error.AztkError(
"The cluster you are trying to connect to does not exist.")
else:
raise
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster_conf = ClusterConfiguration()
cluster_conf.spark_configuration = load_aztk_spark_config()
# read cluster.yaml configuration file, overwrite values with args
file_config, wait = config.read_cluster_config()
cluster_conf.merge(file_config)
if args.size_low_pri is not None:
deprecate("0.9.0", "--size-low-pri has been deprecated.", "Please use --size-low-priority.")
args.size_low_priority = args.size_low_pri
cluster_conf.merge(ClusterConfiguration(
cluster_id=args.cluster_id,
size=args.size,
size_low_priority=args.size_low_priority,
vm_size=args.vm_size,
subnet_id=args.subnet_id,
user_configuration=UserConfiguration(
username=args.username,
password=args.password,
)))
if args.docker_repo and cluster_conf.toolkit:
cluster_conf.toolkit.docker_repo = args.docker_repo
wait = wait if args.wait is None else args.wait
user_configuration = cluster_conf.user_configuration
if user_configuration and user_configuration.username:
ssh_key, password = utils.get_ssh_key_or_prompt(spark_client.secrets_config.ssh_pub_key,
user_configuration.username,
user_configuration.password,
spark_client.secrets_config)
cluster_conf.user_configuration = aztk.spark.models.UserConfiguration(
username=user_configuration.username,
password=password,
ssh_key=ssh_key
)
else:
cluster_conf.user_configuration = None
cluster_conf.validate()
utils.print_cluster_conf(cluster_conf, wait)
with utils.Spinner():
# create spark cluster
cluster = spark_client.create_cluster(
cluster_conf,
wait=wait
)
if wait:
log.info("Cluster %s created successfully.", cluster.id)
else:
log.info("Cluster %s is being provisioned.", cluster.id)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
app_log = spark_client.get_job_application_log(args.job_id, args.app_name)
if args.output:
with utils.Spinner():
with open(os.path.abspath(os.path.expanduser(args.output)),
"w",
encoding="UTF-8") as f:
f.write(app_log.log)
else:
log.print(app_log.log)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster_id = args.cluster_id
cluster = spark_client.get_cluster(cluster_id)
utils.print_cluster(spark_client, cluster)
configuration = spark_client.get_cluster_config(cluster_id)
if configuration and args.show_config:
log.info("-------------------------------------------")
log.info("Cluster configuration:")
utils.print_cluster_conf(configuration, False)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
if args.tail:
utils.stream_logs(client=spark_client,
cluster_id=args.cluster_id,
application_name=args.app_name)
else:
app_logs = spark_client.get_application_log(
cluster_id=args.cluster_id, application_name=args.app_name)
print(app_logs.log)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster_conf = ClusterConfiguration()
cluster_conf.spark_configuration = load_aztk_spark_config()
# read cluster.yaml configuartion file, overwrite values with args
file_config, wait = config.read_cluster_config()
cluster_conf.merge(file_config)
cluster_conf.merge(ClusterConfiguration(
cluster_id=args.cluster_id,
vm_count=args.size,
vm_low_pri_count=args.size_low_pri,
vm_size=args.vm_size,
subnet_id=args.subnet_id,
user_configuration=UserConfiguration(
username=args.username,
password=args.password,
),
docker_repo=args.docker_repo))
wait = wait if args.wait is None else args.wait
user_configuration = cluster_conf.user_configuration
if user_configuration and user_configuration.username:
ssh_key, password = utils.get_ssh_key_or_prompt(spark_client.secrets_config.ssh_pub_key,
user_configuration.username,
user_configuration.password,
spark_client.secrets_config)
cluster_conf.user_configuration = aztk.spark.models.UserConfiguration(
username=user_configuration.username,
password=password,
ssh_key=ssh_key
)
else:
cluster_conf.user_configuration = None
utils.print_cluster_conf(cluster_conf, wait)
spinner = utils.Spinner()
spinner.start()
# create spark cluster
cluster = spark_client.create_cluster(
cluster_conf,
wait=wait
)
spinner.stop()
if wait:
log.info("Cluster %s created successfully.", cluster.id)
else:
log.info("Cluster %s is being provisioned.", cluster.id)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
with utils.Spinner():
copy_output = spark_client.cluster_copy(
cluster_id=args.cluster_id,
source_path=args.source_path,
destination_path=args.dest_path,
internal=args.internal)
[
print_copy_result(node_id, result, err)
for node_id, result, err in copy_output
]
sys.exit(0 if all([result for _, result, _ in copy_output]) else 1)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
if args.tail:
utils.stream_logs(client=spark_client, cluster_id=args.cluster_id, application_name=args.app_name)
else:
app_log = spark_client.cluster.get_application_log(id=args.cluster_id, application_name=args.app_name)
if args.output:
with utils.Spinner():
with open(os.path.abspath(os.path.expanduser(args.output)), "w", encoding="UTF-8") as f:
f.write(app_log.log)
else:
log.print(app_log.log)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
job_conf = JobConfig()
job_conf.merge(args.job_id, args.job_conf)
aztk_applications = []
for application in job_conf.applications:
aztk_applications.append(
aztk.spark.models.ApplicationConfiguration(
name=application.get('name'),
application=application.get('application'),
application_args=application.get('application_args'),
main_class=application.get('main_class'),
jars=[],
py_files=[],
files=[],
driver_java_options=application.get('driver_java_options'),
driver_library_path=application.get('driver_library_path'),
driver_class_path=application.get('driver_class_path'),
driver_memory=application.get('driver_memory'),
executor_memory=application.get('executor_memory'),
driver_cores=application.get('driver_cores'),
executor_cores=application.get('executor_cores')))
# by default, load spark configuration files in .aztk/
spark_configuration = config.load_aztk_spark_config()
# overwrite with values in job_conf if they exist
if job_conf.spark_defaults_conf:
spark_configuration.spark_defaults_conf = job_conf.spark_defaults_conf
if job_conf.spark_env_sh:
spark_configuration.spark_env_sh = job_conf.spark_env_sh
if job_conf.core_site_xml:
spark_configuration.core_site_xml = job_conf.core_site_xml
job_configuration = aztk.spark.models.JobConfiguration(
id=job_conf.id,
applications=aztk_applications,
custom_scripts=job_conf.custom_scripts,
spark_configuration=spark_configuration,
vm_size=job_conf.vm_size,
docker_repo=job_conf.docker_repo,
max_dedicated_nodes=job_conf.max_dedicated_nodes,
max_low_pri_nodes=job_conf.max_low_pri_nodes,
subnet_id=job_conf.subnet_id,
worker_on_master=job_conf.worker_on_master)
#TODO: utils.print_job_conf(job_configuration)
spark_client.submit_job(job_configuration)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
cluster_id = args.cluster_id
if not args.force:
confirmation_cluster_id = input("Please confirm the id of the cluster you wish to delete: ")
if confirmation_cluster_id != cluster_id:
log.error("Confirmation cluster id does not match. Please try again.")
return
if spark_client.delete_cluster(cluster_id):
log.info("Deleting cluster %s", cluster_id)
else:
log.error("Cluster with id '%s' doesn't exist or was already deleted.", cluster_id)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
job_id = args.job_id
if not args.force:
# check if job exists before prompting for confirmation
spark_client.get_job(job_id)
confirmation_cluster_id = input("Please confirm the id of the cluster you wish to delete: ")
if confirmation_cluster_id != job_id:
log.error("Confirmation cluster id does not match. Please try again.")
return
if spark_client.delete_job(job_id):
log.info("Deleting Job %s", job_id)
else:
log.error("Job with id '%s' doesn't exist or was already deleted.", job_id)
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
utils.print_jobs(spark_client.list_jobs())
from azure.batch.models import BatchErrorException
import aztk.spark
import pytest
from aztk.utils import constants
from aztk.error import AztkError
from aztk_cli import config
# base cluster name
dt = datetime.now()
current_time = dt.microsecond
base_cluster_id = "cluster-{}".format(current_time)
# load secrets
# note: this assumes secrets are set up in .aztk/secrets
spark_client = aztk.spark.Client(config.load_aztk_secrets())
# helper method
def wait_until_cluster_deleted(cluster_id: str):
while True:
try:
spark_client.get_cluster(cluster_id)
time.sleep(1)
except AztkError:
# break when the cluster is not found
break
def test_create_cluster():
test_id = "test-create-"
def execute(args: typing.NamedTuple):
spark_client = aztk.spark.Client(config.load_aztk_secrets())
spark_client.stop_job(args.job_id)
print("Stopped Job {0}".format(args.job_id))
def execute(args: typing.NamedTuple):
if not args.wait and args.output:
raise aztk.error.AztkError("--output flag requires --wait flag")
spark_client = aztk.spark.Client(config.load_aztk_secrets())
jars = []
py_files = []
files = []
if args.jars is not None:
jars = args.jars.replace(' ', '').split(',')
if args.py_files is not None:
py_files = args.py_files.replace(' ', '').split(',')
if args.files is not None:
files = args.files.replace(' ', '').split(',')
log.info("-------------------------------------------")
log.info("Spark cluster id: %s", args.cluster_id)
log.info("Spark app name: %s", args.name)
log.info("Wait for app completion: %s", args.wait)
if args.main_class is not None:
log.info("Entry point class: %s", args.main_class)
if jars:
log.info("JARS: %s", jars)
if py_files:
log.info("PY_Files: %s", py_files)
if files:
log.info("Files: %s", files)
if args.driver_java_options is not None:
log.info("Driver java options: %s", args.driver_java_options)
if args.driver_library_path is not None:
log.info("Driver library path: %s", args.driver_library_path)
if args.driver_class_path is not None:
log.info("Driver class path: %s", args.driver_class_path)
if args.driver_memory is not None:
log.info("Driver memory: %s", args.driver_memory)
if args.executor_memory is not None:
log.info("Executor memory: %s", args.executor_memory)
if args.driver_cores is not None:
log.info("Driver cores: %s", args.driver_cores)
if args.executor_cores is not None:
log.info("Executor cores: %s", args.executor_cores)
log.info("Application: %s", args.app)
log.info("Application arguments: %s", args.app_args)
log.info("-------------------------------------------")
spark_client.submit(cluster_id=args.cluster_id,
application=aztk.spark.models.ApplicationConfiguration(
name=args.name,
application=args.app,
application_args=args.app_args,
main_class=args.main_class,
jars=jars,
py_files=py_files,
files=files,
driver_java_options=args.driver_java_options,
driver_library_path=args.driver_library_path,
driver_class_path=args.driver_class_path,
driver_memory=args.driver_memory,
executor_memory=args.executor_memory,
driver_cores=args.driver_cores,
executor_cores=args.executor_cores,
max_retry_count=args.max_retry_count),
remote=args.remote,
wait=False)
if args.wait:
if not args.output:
exit_code = utils.stream_logs(client=spark_client,
cluster_id=args.cluster_id,
application_name=args.name)
else:
with utils.Spinner():
spark_client.wait_until_application_done(
cluster_id=args.cluster_id, task_id=args.name)
application_log = spark_client.get_application_log(
cluster_id=args.cluster_id, application_name=args.name)
with open(os.path.abspath(os.path.expanduser(args.output)),
"w",
encoding="UTF-8") as f:
f.write(application_log.log)
exit_code = application_log.exit_code
sys.exit(exit_code)
