def run(cls):
'''
Main command-line entry point.
:param cls: The class as implicit first argument.
'''
try:
args = _parse_args()
conf = json.loads(args.config_file.read())
# .........................set up logger
Util.get_logger('SPOT', args.log_level)
# .........................check kerberos authentication
if os.getenv('KRB_AUTH'):
kb = Kerberos()
kb.authenticate()
conf['producer'] = {
'bootstrap_servers': [
'{0}:{1}'.format(conf['kafka']['kafka_server'],
conf['kafka']['kafka_port'])
]
}
conf['file_watcher'] = {
'path': conf['pipelines'][args.type]['collector_path'],
'supported_files':
conf['pipelines'][args.type]['supported_files'],
'recursive': True
}
# .........................migrate configs
if not 'local_staging' in conf['pipelines'][args.type].keys():
conf['pipelines'][args.type]['local_staging'] = '/tmp'
if 'max_request_size' in conf['kafka'].keys():
conf['producer']['max_request_size'] = conf['kafka'][
'max_request_size']
if not 'process_opt' in conf['pipelines'][args.type].keys():
conf['pipelines'][args.type]['process_opt'] = ''
if 'recursive' in conf['pipelines'][args.type].keys():
conf['file_watcher']['recursive'] = conf['pipelines'][
args.type]['recursive']
collector = cls(args.type, args.topic, args.skip_conversion,
**conf)
collector.start()
except SystemExit:
raise
except:
sys.excepthook(*sys.exc_info())
sys, exit(1)
def __init__(self):
self._logger = Util.get_logger('SPOT.COMMON.KERBEROS')
principal, keytab, sasl_mech, security_proto = config.kerberos()
if os.getenv('KINITPATH'):
self._kinit = os.getenv('KINITPATH')
else:
self._kinit = "kinit"
self._kinitopts = os.getenv('KINITOPTS')
self._keytab = "-kt {0}".format(keytab)
self._krb_user = principal
if self._kinit == None or self._keytab == None or self._krb_user == None:
self._logger.error(
"Please verify kerberos configuration, some environment variables are missing."
)
sys.exit(1)
if self._kinitopts is None:
self._kinit_cmd = "{0} {1} {2}".format(self._kinit, self._keytab,
self._krb_user)
else:
self._kinit_cmd = "{0} {1} {2} {3}".format(self._kinit,
self._kinitopts,
self._keytab,
self._krb_user)
def start_collector(type, workers_num, id=None):
# generate ingest id
ingest_id = str(datetime.datetime.time(datetime.datetime.now())).replace(
":", "_").replace(".", "_")
# create logger.
logger = Util.get_logger("SPOT.INGEST")
# validate the given configuration exists in ingest_conf.json.
if not type in master_conf["pipelines"]:
logger.error("'{0}' type is not a valid configuration.".format(type))
sys.exit(1)
# validate the type is a valid module.
if not Util.validate_data_source(master_conf["pipelines"][type]["type"]):
logger.error(
"'{0}' type is not configured. Please check you ingest conf file".
format(master_conf["pipelines"][type]["type"]))
sys.exit(1)
# validate if kerberos authentication is required.
if os.getenv('KRB_AUTH'):
kb = Kerberos()
kb.authenticate()
# kafka server info.
logger.info("Initializing kafka instance")
k_server = master_conf["kafka"]['kafka_server']
k_port = master_conf["kafka"]['kafka_port']
# required zookeeper info.
zk_server = master_conf["kafka"]['zookeper_server']
zk_port = master_conf["kafka"]['zookeper_port']
topic = "SPOT-INGEST-{0}_{1}".format(type, ingest_id) if not id else id
kafka = KafkaTopic(topic, k_server, k_port, zk_server, zk_port,
workers_num)
# create a collector instance based on data source type.
logger.info("Starting {0} ingest instance".format(topic))
module = __import__("pipelines.{0}.collector".format(
master_conf["pipelines"][type]["type"]),
fromlist=['Collector'])
# start collector.
ingest_collector = module.Collector(master_conf['hdfs_app_path'], kafka,
type)
ingest_collector.start()
def _initialize_members(self,db_name,hdfs_app_path,kafka_consumer,conf_type,processes):
# get logger instance.
self._logger = Util.get_logger('SPOT.INGEST.WRK.PROXY')
self._db_name = db_name
self._hdfs_app_path = hdfs_app_path
self._kafka_consumer = kafka_consumer
# read proxy configuration.
self._script_path = os.path.dirname(os.path.abspath(__file__))
conf_file = "{0}/ingest_conf.json".format(os.path.dirname(os.path.dirname(self._script_path)))
conf = json.loads(open(conf_file).read())
self._spark_conf = conf["spark-streaming"]
self._conf = conf["pipelines"][conf_type]
self._processes = processes
def _initialize_members(self,db_name,hdfs_app_path,kafka_consumer,conf_type):
# get logger instance.
self._logger = Util.get_logger('SPOT.INGEST.WRK.FLOW')
self._db_name = db_name
self._hdfs_app_path = hdfs_app_path
# read proxy configuration.
self._script_path = os.path.dirname(os.path.abspath(__file__))
conf_file = "{0}/ingest_conf.json".format(os.path.dirname(os.path.dirname(self._script_path)))
conf = json.loads(open(conf_file).read())
self._conf = conf["pipelines"][conf_type]
self._process_opt = self._conf['process_opt']
self._local_staging = self._conf['local_staging']
self.kafka_consumer = kafka_consumer
def _initialize_members(self, db_name, hdfs_app_path, kafka_consumer,
conf_type, processes):
# get logger instance.
self._logger = Util.get_logger('SPOT.INGEST.WRK.PROXY')
self._db_name = db_name
self._hdfs_app_path = hdfs_app_path
self._kafka_consumer = kafka_consumer
# read proxy configuration.
self._script_path = os.path.dirname(os.path.abspath(__file__))
conf_file = "{0}/ingest_conf.json".format(
os.path.dirname(os.path.dirname(self._script_path)))
conf = json.loads(open(conf_file).read())
self._spark_conf = conf["spark-streaming"]
self._conf = conf["pipelines"][conf_type]
self._processes = processes
def start_worker(type, topic, id, processes=None):
logger = Util.get_logger("SPOT.INGEST.WORKER")
# validate the given configuration exists in ingest_conf.json.
if not type in WORKER_CONF["pipelines"]:
logger.error("'{0}' type is not a valid configuration.".format(type))
sys.exit(1)
# validate the type is a valid module.
if not Util.validate_data_source(WORKER_CONF["pipelines"][type]["type"]):
logger.error("The provided data source {0} is not valid".format(type))
sys.exit(1)
# validate if kerberos authentication is requiered.
if os.getenv('KRB_AUTH'):
kb = Kerberos()
kb.authenticate()
# create a worker instance based on the data source type.
module = __import__("pipelines.{0}.worker".format(
WORKER_CONF["pipelines"][type]["type"]),
fromlist=['Worker'])
# kafka server info.
logger.info("Initializing kafka instance")
k_server = WORKER_CONF["kafka"]['kafka_server']
k_port = WORKER_CONF["kafka"]['kafka_port']
# required zookeeper info.
zk_server = WORKER_CONF["kafka"]['zookeper_server']
zk_port = WORKER_CONF["kafka"]['zookeper_port']
topic = topic
# create kafka consumer.
kafka_consumer = KafkaConsumer(topic, k_server, k_port, zk_server, zk_port,
id)
# start worker.
db_name = WORKER_CONF['dbname']
app_path = WORKER_CONF['hdfs_app_path']
ingest_worker = module.Worker(db_name, app_path, kafka_consumer, type,
processes)
ingest_worker.start()
def _initialize_members(self, db_name, hdfs_app_path, kafka_consumer,
conf_type):
# get logger instance.
self._logger = Util.get_logger('SPOT.INGEST.WRK.DNS')
self._db_name = db_name
self._hdfs_app_path = hdfs_app_path
# read proxy configuration.
self._script_path = os.path.dirname(os.path.abspath(__file__))
conf_file = "{0}/ingest_conf.json".format(
os.path.dirname(os.path.dirname(self._script_path)))
conf = json.loads(open(conf_file).read())
self._conf = conf["pipelines"][conf_type]
self._process_opt = self._conf['process_opt']
self._local_staging = self._conf['local_staging']
self.kafka_consumer = kafka_consumer
def start_worker(type,topic,id,processes=None):
logger = Util.get_logger("SPOT.INGEST.WORKER")
# validate the given configuration exists in ingest_conf.json.
if not type in worker_conf["pipelines"]:
logger.error("'{0}' type is not a valid configuration.".format(type));
sys.exit(1)
# validate the type is a valid module.
if not Util.validate_data_source(worker_conf["pipelines"][type]["type"]):
logger.error("The provided data source {0} is not valid".format(type));sys.exit(1)
# validate if kerberos authentication is requiered.
if os.getenv('KRB_AUTH'):
kb = Kerberos()
kb.authenticate()
# create a worker instance based on the data source type.
module = __import__("pipelines.{0}.worker".format(worker_conf["pipelines"][type]["type"]),fromlist=['Worker'])
# kafka server info.
logger.info("Initializing kafka instance")
k_server = worker_conf["kafka"]['kafka_server']
k_port = worker_conf["kafka"]['kafka_port']
# required zookeeper info.
zk_server = worker_conf["kafka"]['zookeper_server']
zk_port = worker_conf["kafka"]['zookeper_port']
topic = topic
# create kafka consumer.
kafka_consumer = KafkaConsumer(topic,k_server,k_port,zk_server,zk_port,id)
# start worker.
db_name = worker_conf['dbname']
app_path = worker_conf['hdfs_app_path']
ingest_worker = module.Worker(db_name,app_path,kafka_consumer,type,processes)
ingest_worker.start()