def main():
config_file = "config.xml"
xml = ET.parse(config_file)
HOST_NAME = xml.find('host_name').text
DB_NAME = xml.find('db_name').text
DISPLAY_ADDR = xml.find("display_addr").text
DISPLAY_PORT = int(xml.find('display_port').text)
ENDPOINT = xml.find("endpoint").text
SECONDS = int(xml.find("seconds").text)
USERNAME = xml.find("username").text
PASSWORD = xml.find("password").text
HOST = 'http://'+ USERNAME+ ':'+ PASSWORD + '@'+ HOST_NAME
print "starting ..."
#print HOST, DB_NAME, DISPLAY_ADDR, DISPLAY_PORT
#readData(HOST, DB_NAME, DISPLAY_ADDR, DISPLAY_PORT)
display = LedDisplay(HOST, DB_NAME, DISPLAY_ADDR, DISPLAY_PORT, ENDPOINT)
# Start the scheduler
sched = Scheduler()
sched.add_interval_job(display.query, seconds=SECONDS )
sched.start()
try:
while True:
time.sleep(0.1)
except KeyboardInterrupt:
print "terminating"
sched.shutdown()
class Timer:
def __init__(self, fieldCanvas, field):
#self.sched = BackgroundScheduler()
self.sched = Scheduler()
self.sched.start()
self.timerValue = 0
self.fieldCanvas = fieldCanvas
self.field = field
def updateTimerValue(self):
self.timerValue += 1
minutes = "0" + str(int(self.timerValue/60)) if int(self.timerValue/60) < 9 else str(int(self.timerValue/60))
secondes = "0" + str(self.timerValue%60) if self.timerValue%60 < 9 else str(self.timerValue%60)
timerString = minutes + ":" + secondes
self.fieldCanvas.itemconfigure(self.field ,text=timerString)
def getTimer(self):
return self.timerValue
def startTimer(self):
#job = self.sched.add_job(self.updateTimerValue, 'interval', seconds = 1)
job = self.sched.add_interval_job(self.updateTimerValue, seconds = 1)
pass
def stopTimer(self):
try:
model.user.setScore(self.getTimer())
self.sched.shutdown(wait=False)
except Exception as e:
#print(e)
pass
class Scheduler(Plugin):
crons = {}
intervals = {}
started = False
def __init__(self):
addEvent("schedule.cron", self.cron)
addEvent("schedule.interval", self.interval)
addEvent("schedule.remove", self.remove)
self.sched = Sched(misfire_grace_time=60)
self.sched.start()
self.started = True
def remove(self, identifier):
for cron_type in ["intervals", "crons"]:
try:
self.sched.unschedule_job(getattr(self, cron_type)[identifier]["job"])
log.debug("%s unscheduled %s", (cron_type.capitalize(), identifier))
except:
pass
def doShutdown(self):
self.stop()
return super(Scheduler, self).doShutdown()
def stop(self):
if self.started:
log.debug("Stopping scheduler")
self.sched.shutdown()
log.debug("Scheduler stopped")
self.started = False
def cron(self, identifier="", handle=None, day="*", hour="*", minute="*"):
log.info('Scheduling "%s", cron: day = %s, hour = %s, minute = %s', (identifier, day, hour, minute))
self.remove(identifier)
self.crons[identifier] = {
"handle": handle,
"day": day,
"hour": hour,
"minute": minute,
"job": self.sched.add_cron_job(handle, day=day, hour=hour, minute=minute),
}
def interval(self, identifier="", handle=None, hours=0, minutes=0, seconds=0):
log.info(
"Scheduling %s, interval: hours = %s, minutes = %s, seconds = %s", (identifier, hours, minutes, seconds)
)
self.remove(identifier)
self.intervals[identifier] = {
"handle": handle,
"hours": hours,
"minutes": minutes,
"seconds": seconds,
"job": self.sched.add_interval_job(handle, hours=hours, minutes=minutes, seconds=seconds),
}
def schedule(file_name, n_jobs, frequency):
'''Schedule the scraper to execute every hour and shut it down after a
certain number of jos have been run'''
# Create a default logger
basicConfig()
# Run the first job
scrape(file_name)
# Instantiate the scheduler
sched = Scheduler()
# Start it
sched.start()
# Schedule the function
sched.add_interval_job(scrape,
args=[file_name],
minutes=frequency,
misfire_grace_time=60)
# Wait to run n_jobs (assuming 1 job per hour, which is 3600 seconds)
sleep(n_jobs * 3600)
# Shutdown the scheduler
sched.shutdown()
def schedule(file_name, n_jobs, frequency):
'''Schedule the scraper to execute every hour and shut it down after a
certain number of jos have been run'''
# Create a default logger
basicConfig()
# Run the first job
scrape(file_name)
# Instantiate the scheduler
sched = Scheduler()
# Start it
sched.start()
# Schedule the function
sched.add_interval_job(scrape, args=[file_name], minutes=frequency,
misfire_grace_time=60)
# Wait to run n_jobs (assuming 1 job per hour, which is 3600 seconds)
sleep(n_jobs * 3600)
# Shutdown the scheduler
sched.shutdown()
def main(config_file):
info = load_config(config_file)
day_of_week = info['days']
hour = info['hour']
minute = info['minute']
config = {
'apscheduler.jobstores.file.class': info['scheduler-config']['class'],
'apscheduler.jobstores.file.path': info['scheduler-config']['path']
}
sched = Scheduler(config)
sched.add_cron_job(run_scraper,
day_of_week=day_of_week,
hour=hour,
minute=minute)
sched.add_cron_job(check_archive, day='first')
sched.start()
print('Press Ctrl+{0} to exit'.format('C'))
try:
# This is here to simulate application activity (which keeps the main thread alive).
while True:
time.sleep(2)
except (KeyboardInterrupt, SystemExit):
sched.shutdown(
) # Not strictly necessary if daemonic mode is enabled but should be done if possibleisched.start()
class SnortScheduler(SimplePlugin):
""" Enables Schduling for Snortmanager """
scheduler = None # The APS instance
def __init__(self, bus):
""" Initiates scheduler. """
SimplePlugin.__init__(self, bus)
self.scheduler = Scheduler()
def __initiate_jobs(self):
""" Adds schedueled tasks if database is empty. """
sched = self.scheduler
sched.add_cron_job(update_snort_rules, hour = 7, jobstore='sql')
sched.add_cron_job(produce_configuration_files, hour = 9, jobstore='sql')
def start(self):
""" Intitates scheduler when Snortmanager starts """
sched = self.scheduler
sched.add_jobstore(ScheduleStore(), 'sql')
if len(sched.get_jobs()) is 0:
self.__initiate_jobs()
sched.start()
def stop(self):
""" Stops Scheduler service when thread dies. """
self.scheduler.shutdown(wait=False)
def restart(self):
""" Restarts the service if necassary. """
self.stop()
self.start()
def get_jobs(self):
return self.scheduler.get_jobs()
class BlueprintHandler:
setter_blacklist = []
getter_blacklist = []
def __init__(self, blueprint, testing=False, testing_count=10):
self.blueprint = blueprint
self.testing = testing
self.testing_count = testing_count
self.scheduler = Scheduler()
def do_step(self):
print "stepping"
try:
# Fetch any outstanding events from the engine process and execute in simulator
while not self.local_queue.empty():
action = self.local_queue.get()
try:
self.blueprint.interface.set(action[0], float(action[1]))
print "Received action:", action
except exceptions.ValueError:
print "Value '" + str(action[1]) + "' is not convertable to float"
points = self.blueprint.interface.get_getters()
self.blueprint.step(stepcount=int(1 / 0.1))
g = {}
for point in points:
if point in BlueprintHandler.getter_blacklist:
continue
g[point] = self.blueprint.interface.get(point)
for k in g.keys():
m = Measurement()
m.bid = self.blueprint.building.buildingID
m.timestamp = datetime.utcnow().replace(tzinfo=utc)
m.uuid = k
m.val = g[k]
m.save()
except:
# print 'error: ', sys.exc_info()
print "trace: ", traceback.print_exc()
def init_scheduler(self):
schedule_store = RAMJobStore()
# Write data every 15 seconds.
job_second = self.scheduler.add_interval_job(self.do_step, 0, 0, 0, 0, 15)
schedule_store.add_job(job_second)
self.scheduler.add_jobstore(schedule_store, "Simulator scheduler", quiet=False)
def start(self, queue=None):
self.local_queue = queue
self.init_scheduler()
self.scheduler.start()
def stop(self):
self.scheduler.shutdown()
class Ddate():
''' Scheduling class to post ddate each morning. '''
def __init__(self, connection, target):
''' Set irc details, initialise scheduler, set daily task,
and grab the current date. '''
self.con = connection
self.target = target
self.sched = Scheduler()
self.sched.start()
self.fetch_ddate()
# scheduler for 0915 (server is gmt+1)
self.sched.add_cron_job(self.change_ddate, hour=10, minute=15)
def change_ddate(self):
''' update ddate and announce to channel. '''
self.fetch_ddate()
self.post_ddate()
def fetch_ddate(self):
''' fetch ddate from shell.'''
date = check_output(['ddate'])
# trim `today is ` and `\n`
self.ddate = date[9:-2]
def post_ddate(self):
''' post date to connection/channel supplied on startup. '''
self.con.send_msg(self.target, self.ddate)
def __del__(self):
''' Kill scheduler on close. '''
self.sched.shutdown()
def main(config_file):
info = load_config(config_file)
day_of_week = info['days']
hour = info['hour']
minute = info['minute']
config = {
'apscheduler.jobstores.file.class': info['scheduler-config']['class'],
'apscheduler.jobstores.file.path': info['scheduler-config']['path']
}
sched = Scheduler(config)
sched.add_cron_job(run_scraper, day_of_week=day_of_week, hour=hour, minute=minute)
sched.add_cron_job(check_archive, day='first')
sched.start()
print('Press Ctrl+{0} to exit'.format('C'))
try:
# This is here to simulate application activity (which keeps the main thread alive).
while True:
time.sleep(2)
except (KeyboardInterrupt, SystemExit):
sched.shutdown() # Not strictly necessary if daemonic mode is enabled but should be done if possibleisched.start()
def main():
sched = Scheduler()
sched.start()
c_date = datetime.today()
date_ = '2013-11-30'
delay = timedelta(minutes=2)
all_portfolios = combine_lists(PROGRESS_PORTFOLIOS,FIS_GROUP_PORTFOLIOS)
# end of week jobs
#-----------------
if c_date.weekday == 4:
# runs at 6pm Friday evening
sched_date = datetime(c_date.year, c_date.month, c_date.day, 18, 0, 0)
sched.add_date_job(
axys_job,
sched_date,
[MODEL_PORTFOLIOS,all_portfolios,date_])
# monthly jobs
#-------------
if c_date.day == 1:
# runs at 10 am
sched_date = datetime(c_date.year, c_date.month, c_date.day, 10, 0, 0)
sched.add_date_job(
axys_job,
sched_date,
[MODEL_PORTFOLIOS,all_portfolios,date_])
sched_date = sched_date + delay
# keep script 'running' in order to allow the scheduler to stay open and run
# the jobs added
time.sleep(60)
sched.shutdown()
def main():
from apscheduler.scheduler import Scheduler
if len(sys.argv) != 2:
sys.exit("usage: %s <config-file>" % sys.argv[0])
try:
from ConfigParser import ConfigParser
except ImportError: # python3
from configparser import ConfigParser
try:
config = ConfigParser(inline_comment_prefixes=(';',))
except TypeError: # not python3
config = ConfigParser()
config.readfp(open(sys.argv[1]))
global logfile
logfile = config.get("global", "logfile")
FORMAT = "%(asctime)-15s: %(message)s"
logging.basicConfig(level=logging.INFO, filename=logfile, filemode='w',
format=FORMAT)
# Set time on WDLXTV systems
rdate = "/usr/sbin/rdate"
if os.path.exists(rdate) and os.access(rdate, os.X_OK):
cmd = [rdate, "ntp.internode.on.net"]
subprocess.Popen(cmd).wait()
logging.info("Main process PID: %d, use this for sending SIGHUP "
"for re-reading the schedule-file", os.getpid())
global tuners
tuners = TUNERS(config.get("global", "tuners"))
global hdhomerun_config
hdhomerun_config = config.get("global", "hdhomerun_config")
schedule_file = config.get("global", "schedule_file")
media_dir = config.get("global", "media_dir")
channelmap = {}
for opt in config.options("channelmap"):
channelmap[opt] = config.get("channelmap", opt).split(",")
while True:
global reload_jobs, shutdown
reload_jobs = False
shutdown = False
sched = Scheduler(misfire_grace_time=60, daemonic=False)
sched.start()
signal.signal(signal.SIGHUP, sighup_handler)
signal.signal(signal.SIGTERM, sigterm_handler)
schedule_jobs(sched, schedule_file, channelmap, media_dir)
while not (reload_jobs or shutdown):
signal.pause()
sched.shutdown()
if shutdown:
sys.exit(0)
class Scheduler(Plugin):
crons = {}
intervals = {}
started = False
def __init__(self):
addEvent('schedule.cron', self.cron)
addEvent('schedule.interval', self.interval)
addEvent('schedule.remove', self.remove)
self.sched = Sched(misfire_grace_time = 60)
self.sched.start()
self.started = True
def remove(self, identifier):
for cron_type in ['intervals', 'crons']:
try:
self.sched.unschedule_job(getattr(self, cron_type)[identifier]['job'])
log.debug('%s unscheduled %s', (cron_type.capitalize(), identifier))
except:
pass
def doShutdown(self):
super(Scheduler, self).doShutdown()
self.stop()
def stop(self):
if self.started:
log.debug('Stopping scheduler')
self.sched.shutdown()
log.debug('Scheduler stopped')
self.started = False
def cron(self, identifier = '', handle = None, day = '*', hour = '*', minute = '*'):
log.info('Scheduling "%s", cron: day = %s, hour = %s, minute = %s', (identifier, day, hour, minute))
self.remove(identifier)
self.crons[identifier] = {
'handle': handle,
'day': day,
'hour': hour,
'minute': minute,
'job': self.sched.add_cron_job(handle, day = day, hour = hour, minute = minute)
}
def interval(self, identifier = '', handle = None, hours = 0, minutes = 0, seconds = 0):
log.info('Scheduling %s, interval: hours = %s, minutes = %s, seconds = %s', (identifier, hours, minutes, seconds))
self.remove(identifier)
self.intervals[identifier] = {
'handle': handle,
'hours': hours,
'minutes': minutes,
'seconds': seconds,
'job': self.sched.add_interval_job(handle, hours = hours, minutes = minutes, seconds = seconds)
}
class Main(Daemon):
"""
do some things
"""
def __init__(self, pidfile, cfgfile):
Daemon.__init__(self, pidfile)
self.jobs = {}
self.immediately = False
self.scheduler = Scheduler(daemonic=False)
self.logger = logging.getLogger(self.__class__.__name__)
if os.path.exists(cfgfile):
with open(cfgfile, 'rt') as f:
config = yaml.load(f.read())
for k1 in config.keys():
if k1 == 'version':
pass
if k1 == 'immediately':
self.immediately = config[k1]
elif k1 == 'taobao':
self.jobs[k1] = config[k1]
self.jobs[k1]['id'] = None
if 'chktime' in self.jobs[k1].keys():
self.jobs[k1]['btime'] = time.strptime(self.jobs[k1]['chktime'].split('-')[0], '%H:%M')
self.jobs[k1]['etime'] = time.strptime(self.jobs[k1]['chktime'].split('-')[1], '%H:%M')
if self.jobs[k1]['btime'] >= self.jobs[k1]['etime']:
raise ValueError('"chktime" is illegal')
else:
raise ValueError('There is no "chktime" be found in configure.')
else:
pass
else:
self.logger.error('{0} not found'.format(cfgfile))
def job_main(self):
st_beg = self.jobs['taobao']['btime']
st_end = self.jobs['taobao']['etime']
dt_beg = datetime.now().replace(hour=st_beg.tm_hour, minute=st_beg.tm_min)
dt_end = datetime.now().replace(hour=st_end.tm_hour, minute=st_end.tm_min)
td_rnd = dt_end - dt_beg
dt_rnd = dt_beg + timedelta(seconds=randint(1, td_rnd.days * 86400 + td_rnd.seconds - 1))
if dt_rnd <= datetime.now():
dt_rnd += timedelta(days=1)
self.jobs['taobao']['id'] = self.scheduler.add_date_job(lambda: self.job_taobao(), dt_rnd)
def job_taobao(self):
for v in self.jobs['taobao']['account']:
taobao = Taobao(v['username'], v['password'])
if taobao.login():
taobao.checkin()
def run(self):
if self.immediately:
self.job_taobao()
self.immediately = False
self.scheduler.add_cron_job(lambda: self.job_main(), hour='0', minute='1')
self.scheduler.start()
stopevent.wait()
self.scheduler.shutdown()
class Application(object):
def __init__(self):
self.books = []
self.logger = logging.getLogger(__name__)
self.scheduler = Scheduler()
self.scheduler.add_interval_job(self.process_book_list, seconds=30)
self.store = AmazonStore()
self.notifier = NotifyMyAndroid(bookworm.NOTIFICATION_KEY)
self.load()
def load(self):
if os.path.exists('books.dat'):
with open('books.dat', 'rb') as f:
self.books = pickle.load(f)
self.logger.debug('Loaded {0} books'.format(len(self.books)))
def save(self):
if(len(self.books) > 0 ):
with open('books.dat', 'wb') as f:
self.logger.debug('Saving books')
pickle.dump(self.books, f, -1)
def start(self):
self.logger.info('Starting application')
self.scheduler.start()
def stop(self):
self.logger.info('Stopping application')
self.save()
self.scheduler.shutdown()
def process_book_list(self):
self.logger.info('Processing book list')
for book in self.books:
self.logger.debug('Updating price for {0}'.format(book.title))
latestPrice = self.store.get_book_price(book.isbn)
self.logger.debug('latest: {0}, saved: {1}'.format(latestPrice, book.current_price))
if latestPrice != None and latestPrice != book.current_price:
if latestPrice < book.current_price:
self.logger.debug('Found a lower price!')
self.notifier.send_notification(self.build_notification_message(latestPrice, book.title, self.store.get_product_url(book.isbn)), 'Price Update')
book.current_price = latestPrice
def build_notification_message(self, price, title, productUrl):
return (u'{0} is now £{1}\n{2}'.format(title, price, productUrl))
def get_resolver(self):
return self.store.get_book_details
def register_new_book(self, book):
self.books.append(book)
class Controller():
def __init__(self):
self.configList = []
self.scheduler = Scheduler()
self.scheduler.start()
self.uploadConfigList()
self.mqttClient = mqtt.Client(mqttClientName)
self.mqttClient.on_connect = onConnect
self.mqttClient.on_disconnect = onDisconnect
self.mqttClient.on_log = onLog
self.mqttClient.on_publish = onPublish
self.mqttClient.on_message = onMessage
self.mqttClient.connect(mqttBrokerHost)
self.mqttClient.loop_start()
def stop(self):
self.scheduler.shutdown()
self.mqttClient.loop_stop()
self.mqttClient.disconnect()
def uploadConfigList(self):
tableRows = None
try:
tableRows = getTableRows(scheduleTable)
except sqlite3.Error as err:
logger.error('SQLite: error: ' + str(err))
self.configList.clear()
for row in tableRows:
config = Config(topic=row[0],
weekday=row[1],
hour=row[2],
minute=row[3],
second=row[4])
logger.info('Uploaded config list with: ' + str(config.topic) +
' ' + str(config.weekday) + ' ' + str(config.hour) +
' ' + str(config.minute) + ' ' + str(config.second))
self.configList.append(config)
self.updateScheduler()
def updateScheduler(self):
for config in self.configList:
self.scheduler.add_cron_job(func=self.requestUpdate,
args=[config.topic],
day_of_week=config.weekday,
hour=config.hour,
minute=config.minute,
second=config.second)
def requestUpdate(self, mqttTopic):
self.mqttClient.publish(mqttTopic, mqttUpdateTrigger, int(mqttQos))
class HeartBeat(Singleton):
def __init__(self):
self.scheduler = Scheduler()
credentials = pika.PlainCredentials('inspur', 'inspur')
flag, rabbitmq_ip = utils.getRabbitmqIp()
# 这里可以连接远程IP，请记得打开远程端口
parameters = pika.ConnectionParameters(rabbitmq_ip, 5672, '/',
credentials)
self.connection = pika.BlockingConnection(parameters)
self.channel = self.connection.channel()
"""
heartbeat task
"""
def agentHeartBeat(self):
logger.debug("agent heart beat begin!")
self.scheduler.add_interval_job(self.heartBeatWithServer, seconds=3)
self.scheduler.start()
print('Press Ctrl+{0} to exit'.format('Break' if os.name ==
'nt' else 'C'))
try:
while True:
if not utils.httpdstate:
sys.exit(0)
except KeyboardInterrupt:
self.scheduler.shutdown()
finally:
self.connection.close()
def heartBeatWithServer(self):
agentInfo = {'type': constr.NODE_STATUS, 'token': constr.NODE_TOKEN}
flag, agentid = utils.getAgentId()
if not (flag and agentid):
return
agentInfo['id'] = agentid
logger.debug("node hb info : {}".format(str(agentInfo)))
try:
self.channel.basic_publish(exchange='exchangeTest',
routing_key='heartBeatKey',
body=json.dumps(agentInfo))
except Exception, e:
logger.error("heartbeat exception: {}".format(e.message))
credentials = pika.PlainCredentials('inspur', 'inspur')
flag, rabbitmq_ip = utils.getRabbitmqIp()
# 这里可以连接远程IP，请记得打开远程端口
parameters = pika.ConnectionParameters(rabbitmq_ip, 5672, '/',
credentials)
self.connection = pika.BlockingConnection(parameters)
self.channel = self.connection.channel()
pass
return True
def main():
in_daemon_mode = conf.alarm_send_in_daemon_mode
if not in_daemon_mode:
check()
else:
from apscheduler.scheduler import Scheduler
minuteScheduler = Scheduler()
sleep_seconds = 60 # just 60 seconds
minuteScheduler.add_interval_job(check, seconds=sleep_seconds)
minuteScheduler.start()
while 1:
time.sleep(9999)
minuteScheduler.shutdown()
def main():
in_daemon_mode = conf.alarm_send_in_daemon_mode
if not in_daemon_mode:
check()
else:
from apscheduler.scheduler import Scheduler
minuteScheduler = Scheduler()
sleep_seconds = 60 # just 60 seconds
minuteScheduler.add_interval_job(check, seconds=sleep_seconds)
minuteScheduler.start()
while 1:
time.sleep(9999)
minuteScheduler.shutdown()
def run(self):
scheduler = Scheduler()
if self.config["SCHEDULE"].get("packages") is not None:
scheduler.add_interval_job(self.processor.process, **self.config["SCHEDULE"]["packages"])
scheduler.start()
try:
while True:
time.sleep(999)
except KeyboardInterrupt:
logger.info("Shutting down Carrier...")
scheduler.shutdown(wait=False)
class TestStandaloneScheduler(object):
def setup(self):
self.scheduler = Scheduler(standalone=True)
self.scheduler.add_date_job(lambda: None,
datetime.now() + timedelta(1))
self.thread = Thread(target=self.scheduler.start)
self.thread.start()
def teardown(self):
self.scheduler.shutdown(True)
self.thread.join(3)
def test_scheduler_running(self):
time.sleep(.1)
eq_(self.scheduler.running, True)
def run(self):
scheduler = Scheduler()
if self.config["SCHEDULE"].get("packages") is not None:
scheduler.add_interval_job(self.processor.process,
**self.config["SCHEDULE"]["packages"])
scheduler.start()
try:
while True:
time.sleep(999)
except KeyboardInterrupt:
logger.info("Shutting down Carrier...")
scheduler.shutdown(wait=False)
def scheduleNotification(username, password, receivers, subject, message, attachments, timestring):
logging.basicConfig()
scheduler = Scheduler()
scheduler.start()
sentOn = datetime.datetime.strptime(timestring,"%Y-%m-%dT%H:%M")
scheduler.add_date_job(emailUser,sentOn,[username,password,receivers.split(","),subject,message,attachments])
atexit.register(lambda:scheduler.shutdown(wait=False))
def last_tag_update():
cron = Scheduler(daemon=True)
cron.start()
@cron.interval_schedule(minutes=2)
def job_function():
last_tag_get_update()
atexit.register(lambda: cron.shutdown(wait=False))
def scheduleEmailListener(username,password,response_subject,automated_response):
logging.basicConfig()
scheduler=Scheduler()
scheduler.start()
scheduler.add_interval_job(respondToEmails,minutes=1,args=[username,password,response_subject,automated_response])
mongo.deleteEntry("fields","main",{})
mongo.addEntry("fields","main",{"email":username,"password":password})
atexit.register(lambda:scheduler.shutdown(wait=False))
return scheduler
def create_app(object_name):
app = Flask(__name__)
app.config.from_object(object_name)
cors = CORS(app,
resources={r"/*": {
"origins": "*"
}},
allow_headers=['Authorization', 'Content-Type'])
db.init_app(app)
cron = Scheduler(daemon=True)
cron.start()
@app.route('/')
def index():
return render_template("index.html")
# Importing Blueprints
from controllers.cas import blueprint_cas
# Registering blueprints
app.register_blueprint(blueprint_cas)
# Custom HTTP error handlers
@app.errorhandler(400)
def custom_400(error):
print error
return jsonify(message=error.description['message']), 400
@app.errorhandler(401)
def custom_401(error):
return jsonify(message=error.description['message']), 401
@app.errorhandler(403)
def custom_403(error):
return jsonify(message=error.description['message']), 403
@app.errorhandler(404)
def custom_404(error):
return jsonify(message="Item or resource not found"), 404
@app.errorhandler(405)
def custom_405(error):
return jsonify(message="Not allowed"), 405
@app.errorhandler(500)
def custom_500(error):
return jsonify(message=error.description['message']), 500
#@app.errorhandler(Exception)
#def unhandled_exception(e):
# return jsonify(message=str(e)),500
# Shutdown your cron thread if the web process is stopped
atexit.register(lambda: cron.shutdown(wait=False))
return app
def main():
from apscheduler.scheduler import Scheduler
from ConfigParser import ConfigParser
config = ConfigParser()
config.readfp(open('config-file'))
global logfile
logfile = config.get("global", "logfile")
FORMAT = "%(asctime)-15s: %(message)s"
logging.basicConfig(level=logging.INFO, filename=logfile, filemode='w',
format=FORMAT)
# Set time on WDLXTV systems
rdate = "/usr/sbin/rdate"
if os.path.exists(rdate) and os.access(rdate, os.X_OK):
cmd = [rdate, "ntp.internode.on.net"]
subprocess.Popen(cmd).wait()
logging.info("Main process PID: %d, use this for sending SIGHUP "
"for re-reading the schedule-file", os.getpid())
global tuners
tuners = TUNERS(config.get("global", "tuners"))
global hdhomerun_config
hdhomerun_config = config.get("global", "hdhomerun_config")
schedule_file = config.get("global", "schedule_file")
media_dir = config.get("global", "media_dir")
channelmap = {}
for opt in config.options("channelmap"):
channelmap[opt] = config.get("channelmap", opt).split(",")
while True:
global reload_jobs
reload_jobs = False
sched = Scheduler(misfire_grace_time=60, daemonic=False)
sched.start()
signal.signal(signal.SIGHUP, sighup_handler)
schedule_jobs(sched, schedule_file, channelmap, media_dir)
while not reload_jobs:
signal.pause()
sched.shutdown()
def go():
sched = Scheduler()
# startTime = datetime.datetime.now()+datetime.timedelta(seconds=60*60)
# startTime = datetime.datetime.now()+datetime.timedelta(seconds=60*15)
# startTime = datetime.datetime.now()+datetime.timedelta(seconds=60*5)
startTime = datetime.datetime.now()+datetime.timedelta(seconds=20)
scheduleJobs(sched, startTime)
sched.start()
while runStatus.run:
time.sleep(0.1)
print("Scraper stopping scheduler")
sched.shutdown()
nt.dirNameProxy.stop()
class ActivityScheduler(baseRunner.BaseRunner):
def __init__(self):
super(ActivityScheduler, self).__init__()
# set logging options as defined in config file
logConf = self.config._sections["logging"]
# remove default __name__ item
del logConf["__name__"]
logConf["level"] = int(logConf["level"])
logConf["filename"] = join(dirname(__file__), logConf["filename"])
logging.basicConfig(**logConf)
# initialize scheduler
self.scheduler = Scheduler()
self.scheduler.start()
# create initial schedule
if not self.scheduler.get_jobs():
self.createSchedule()
# main loop
while True:
try:
time.sleep(10)
except KeyboardInterrupt:
logging.info("Shutting down..")
self.scheduler.shutdown()
break
def createSchedule(self):
logging.info("Schedule requests..")
schedules = self.config._sections["schedule"]
# remove default __name__ item
del schedules["__name__"]
for methodName, schedule in schedules.items():
# schedule handler requests (wrapper method gets called with
# cron-like notation and the method name)
# name parameter is given for logging/debugging purposes only
self.scheduler.add_cron_job(self.wrap, *schedule.split(), \
args=[methodName], misfire_grace_time=120, name=methodName)
class ActivityScheduler(baseRunner.BaseRunner):
def __init__(self):
super(ActivityScheduler, self).__init__()
# set logging options as defined in config file
logConf = self.config._sections["logging"]
# remove default __name__ item
del logConf["__name__"]
logConf["level"] = int(logConf["level"])
logConf["filename"] = join(dirname(__file__), logConf["filename"])
logging.basicConfig(**logConf)
# initialize scheduler
self.scheduler = Scheduler()
self.scheduler.start()
# create initial schedule
if not self.scheduler.get_jobs():
self.createSchedule()
# main loop
while True:
try:
time.sleep(10)
except KeyboardInterrupt:
logging.info("Shutting down..")
self.scheduler.shutdown()
break
def createSchedule(self):
logging.info("Schedule requests..")
schedules = self.config._sections["schedule"]
# remove default __name__ item
del schedules["__name__"]
for methodName, schedule in schedules.items():
# schedule handler requests (wrapper method gets called with
# cron-like notation and the method name)
# name parameter is given for logging/debugging purposes only
self.scheduler.add_cron_job(self.wrap, *schedule.split(), \
args=[methodName], misfire_grace_time=120, name=methodName)
class CronJob(HTPCObject):
# List
_notifications_to_publish = []
def __init__(self, name, schedule):
super(CronJob, self).__init__(name)
# Configure scheduler
self.scheduler = Scheduler(schedule)
day, hour, minute = schedule
self.scheduler.add_cron_job(self.run, day=day, hour=hour, minute=minute)
# Initialize notifications to offer
for notification in self._notifications_to_publish:
self._published_notifications.append(notification)
def start(self):
self.scheduler.start()
def stop(self):
self.scheduler.shutdown()
def run(self):
raise NotImplementedError("I don't know how to run the cron job!")
def home(request):
global sched
# if pastebin button is pressed, invoke masterScrape to scrape pastebin archive
# scheduler interval of 3 minutes
if 'run_scraper_pastebin' in request.POST:
logging.basicConfig()
sched = Scheduler(standalone=True)
args = ['http://pastebin.com/archive', '1', '180']
call_command('masterScrape', *args)
def scrape_sched():
call_command('masterScrape', *args)
sched.add_interval_job(scrape_sched, seconds=180, max_instances=1000)
sched.start()
# if pastie button is pressed, invoke masterScrape to scrape pastie archive
# scheduler interval of 10 minutes
if 'run_scraper_pastie' in request.POST:
logging.basicConfig()
sched = Scheduler(standalone=True)
args = ['http://pastie.org/pastes', '0', '60']
call_command('masterScrape', *args)
def scrape_sched_pastie():
call_command('masterScrape', *args)
sched.add_interval_job(scrape_sched_pastie,
seconds=60,
max_instances=1000)
sched.start()
# sets flags to stop scrape function
if 'stop_scraper' in request.POST:
args = ['stop']
call_command('masterScrape', *args)
sched.shutdown(wait=False)
return render(request, 'scraper/index2.0.html')
def home(request):
global sched
# if pastebin button is pressed, invoke masterScrape to scrape pastebin archive
# scheduler interval of 3 minutes
if 'run_scraper_pastebin' in request.POST:
logging.basicConfig()
sched = Scheduler(standalone = True, coalesce = True)
#coalesce is used to avoid APS scheduler missed job warning which in turn causes crashing
args = ['http://pastebin.com/archive', '1', '180']
call_command('masterScrape', *args)
def scrape_sched():
call_command('masterScrape', *args)
sched.add_interval_job(scrape_sched, seconds = 180, max_instances = 1000)
sched.start()
# if pastie button is pressed, invoke masterScrape to scrape pastie archive
# scheduler interval of 10 minutes
if 'run_scraper_pastie' in request.POST:
logging.basicConfig()
sched = Scheduler(standalone = True, coalesce = True)
args = ['http://pastie.org/pastes', '0', '60']
call_command('masterScrape', *args)
def scrape_sched_pastie():
call_command('masterScrape', *args)
sched.add_interval_job(scrape_sched_pastie, seconds = 60, max_instances = 1000)
sched.start()
# sets flags to stop scrape function
if 'stop_scraper' in request.POST:
args = ['stop']
call_command('masterScrape', *args)
sched.shutdown(wait = False)
return render(request, 'scraper/index2.0.html')
class MyScheduler:
EVENTS = {
'1': 'EVENT_SCHEDULER_START',
'2': 'EVENT_SCHEDULER_SHUTDOWN',
'3': 'EVENT_JOBSTORE_ADDED',
'4': 'EVENT_JOBSTORE_REMOVED',
'5': 'EVENT_JOBSTORE_JOB_ADDED',
'32': 'EVENT_JOBSTORE_JOB_REMOVED',
'64': 'EVENT_JOB_EXECUTED',
'128': 'EVENT_JOB_ERROR',
'256': 'EVENT_JOB_MISSED'
}
def __init__(self, db_path='sqlite:///scheduler.db'):
self.scheduler = Scheduler()
self.scheduler.add_jobstore(SQLAlchemyJobStore(url=db_path), 'default')
def start(self):
self.scheduler.start()
def add_job(self, job, date, args):
job = self.scheduler.add_date_job(job, date, args)
print job
def jobs(self):
return self.scheduler.get_jobs()
def remove_job(self, notfication_id):
jobs = self.jobs()
for job in jobs:
if int(job.args[0]) == int(notfication_id):
self.scheduler.unschedule_job(job)
return True
return False
def shutdown(self):
self.scheduler.shutdown()
def init_crawler():
global link_re, script_re, conn, cur, sched, first_run
conn = sqlite3.connect('steam_summer_sale.db', check_same_thread=False)
cur = conn.cursor()
sql_f = open('schema.sql')
cur.executescript(sql_f.read())
script_re = re.compile("InitDailyDealTimer\( \$\('\w+'\), ([0-9]+) \);")
link_re = re.compile('http:\/\/store\.steampowered\.com\/(app|sub)\/([0-9]+)')
sched = Scheduler()
sched.add_interval_job(do_job, minutes=2)
try:
first_run = True
do_job()
first_run = False
sched.start()
while True:
time.sleep(300)
except:
sched.shutdown()
conn.close()
def __init__(self, socketio):
self.main = Main(socketio)
self.execute_now()
cron = Scheduler(daemon=True)
cron.start()
@cron.interval_schedule(hours=1)
def job_function():
# Do your work here
self.main.start()
# Shutdown your cron thread if the web process is stopped
atexit.register(lambda: cron.shutdown(wait=False))
def start_schedule():
#if __name__ == '__main__':
scheduler_pl = Scheduler(daemonic = False)
scheduler_pl.print_jobs()
scheduler_pl.shutdown()
scheduler_pl.add_jobstore(ShelveJobStore('/tmp/db_pl_schedule'), 'file')
v_current_jobs = scheduler_pl.get_jobs()
print v_current_jobs
if v_current_jobs: # 如果job存在的话,先请客
scheduler_pl.unschedule_func(upload_processlist)
scheduler_pl.add_interval_job(upload_processlist, minutes=1)
scheduler_pl.start()
print 'success!'
scheduler_pl.print_jobs()
'''
class JobDaemon(object):
def __init__(self):
# Start the scheduler
self.sched = Scheduler(daemonic=True)
# register three exit handlers
atexit.register(lambda: self.sched.shutdown())
self.sched.start()
def job_daemon_execute(self, module_name, func_name, *args, **kwargs):
print 'Client request comming... ',
print '[Module name:', module_name, '/',
print 'Function name:', func_name, ']'
command_module = importlib.import_module(module_name)
func = getattr(command_module, func_name)
result = func(sched=self.sched, *args, **kwargs)
return result
def task():
cron = Scheduler(daemon=True)
cron.start()
@cron.interval_schedule(minutes=1)
def job_function():
all_user = Steemit_User().query.all()
for user in all_user:
analysis_self = Analiysis().query.filter_by(
steemit_user_id=user.id)
end_date = analysis_self.all()[-1].end_date
week_control = (datetime.now() - end_date).total_seconds()
if week_control / 86400 > 6.99:
status, response = get_url(user.steem_name)
analiz_create(response, user, user.steem_name, end_date)
atexit.register(lambda: cron.shutdown(wait=False))
def main():
d = date.today()
t = time(00,30)
startDate = datetime.combine(d,t)
startDate = startDate.replace(tzinfo=tz.tzutc())
startDate = startDate.astimezone(tz.tzlocal())
timeStr = startDate.strftime("%Y-%m-%d %H:%M:%S")
print timeStr
startDate = datetime.strptime(timeStr, "%Y-%m-%d %H:%M:%S")
if (datetime.now() - startDate) > timedelta(seconds=1):
startDate = startDate + timedelta(days=1)
sched = Scheduler()
sched.start()
sched.add_interval_job(runDateDump, days=1, start_date=startDate.strftime("%Y-%m-%d %H:%M:%S"))
sched.print_jobs()
while True:
sleep(1)
print "."
atexit.register(lambda: sched.shutdown(wait=True))
class TestRunningScheduler(object):
def setup(self):
self.scheduler = Scheduler()
self.scheduler.start()
def teardown(self):
if self.scheduler.running:
self.scheduler.shutdown()
def test_shutdown_timeout(self):
self.scheduler.shutdown()
@raises(SchedulerAlreadyRunningError)
def test_scheduler_double_start(self):
self.scheduler.start()
@raises(SchedulerAlreadyRunningError)
def test_scheduler_configure_running(self):
self.scheduler.configure({})
def test_scheduler_double_shutdown(self):
self.scheduler.shutdown()
self.scheduler.shutdown(False)
class TestRunningScheduler(object):
def setup(self):
self.scheduler = Scheduler()
self.scheduler.start()
def teardown(self):
if self.scheduler.running:
self.scheduler.shutdown()
def test_shutdown_timeout(self):
self.scheduler.shutdown()
@raises(SchedulerAlreadyRunningError)
def test_scheduler_double_start(self):
self.scheduler.start()
@raises(SchedulerAlreadyRunningError)
def test_scheduler_configure_running(self):
self.scheduler.configure({})
def test_scheduler_double_shutdown(self):
self.scheduler.shutdown()
self.scheduler.shutdown(False)
status = 500
print("Exception:", e)
if request.data:
response = {"status": status, "body": "Internal Server Error"}
r = Response(
response=json.dumps(response),
status=status,
mimetype="application/json",
)
else:
r = render_template("500.html", msg=e)
return r
@app.errorhandler(404)
def not_found(e):
return render_template("404.html")
@cron.interval_schedule(minutes=10)
def job_function():
for folder in os.listdir(app.config["UPLOAD_FOLDER"]):
if folder not in IGNORE_FILES:
folder_path = os.path.join(app.config["UPLOAD_FOLDER"], folder)
shutil.rmtree(folder_path)
atexit.register(lambda: cron.shutdown(wait=False))
elif (event.retval == False):
set_status(event.job.name, 4)
logger.error("Event #" + str(event.job.name) + " had an error")
if event.exception:
print event.exception
logger.fatal("Event #" + str(event.job.name) + ' - job crashed :(')
#start the scheduler
sched = Scheduler()
sched.add_jobstore(WriteBackShelveJobStore('jobstore.db'), 'shelve')
sched.start()
sched.add_listener(my_listener, EVENT_JOB_EXECUTED | EVENT_JOB_ERROR)
process_events()
# Process events list every 10 seconds
sched.add_interval_job(process_events, seconds=10)
# Remove completed events from db every minute
sched.add_interval_job(remove_events, minutes=1)
print "Dispatcher started..."
try:
while True:
time.sleep(1)
finally:
sched.shutdown()
class AlertSchedulerHandler():
FILENAME = 'definitions.json'
TYPE_PORT = 'PORT'
TYPE_METRIC = 'METRIC'
TYPE_SCRIPT = 'SCRIPT'
TYPE_WEB = 'WEB'
TYPE_RECOVERY = 'RECOVERY'
def __init__(self, cachedir, stacks_dir, common_services_dir, host_scripts_dir,
alert_grace_period, cluster_configuration, config, recovery_manager,
in_minutes=True):
self.cachedir = cachedir
self.stacks_dir = stacks_dir
self.common_services_dir = common_services_dir
self.host_scripts_dir = host_scripts_dir
self._cluster_configuration = cluster_configuration
if not os.path.exists(cachedir):
try:
os.makedirs(cachedir)
except:
logger.critical("[AlertScheduler] Could not create the cache directory {0}".format(cachedir))
self.APS_CONFIG = {
'apscheduler.threadpool.core_threads': 3,
'apscheduler.coalesce': True,
'apscheduler.standalone': False,
'apscheduler.misfire_grace_time': alert_grace_period
}
self._collector = AlertCollector()
self.__scheduler = Scheduler(self.APS_CONFIG)
self.__in_minutes = in_minutes
self.config = config
self.recovery_manger = recovery_manager
# register python exit handler
ExitHelper().register(self.exit_handler)
def exit_handler(self):
"""
Exit handler
"""
self.stop()
def update_definitions(self, heartbeat):
"""
Updates the persisted alert definitions JSON.
:param heartbeat:
:return:
"""
if 'alertDefinitionCommands' not in heartbeat:
logger.warning("There are no alert definition commands in the heartbeat; unable to update definitions")
return
# prune out things we don't want to store
alert_definitions = []
for command in heartbeat['alertDefinitionCommands']:
command_copy = command.copy()
# no need to store these since we always use the in-memory cached values
if 'configurations' in command_copy:
del command_copy['configurations']
alert_definitions.append(command_copy)
# write out the new definitions
with open(os.path.join(self.cachedir, self.FILENAME), 'w') as f:
json.dump(alert_definitions, f, indent=2)
# reschedule only the jobs that have changed
self.reschedule()
def __make_function(self, alert_def):
return lambda: alert_def.collect()
def start(self):
""" loads definitions from file and starts the scheduler """
if self.__scheduler is None:
return
if self.__scheduler.running:
self.__scheduler.shutdown(wait=False)
self.__scheduler = Scheduler(self.APS_CONFIG)
alert_callables = self.__load_definitions()
# schedule each definition
for _callable in alert_callables:
self.schedule_definition(_callable)
logger.info("[AlertScheduler] Starting {0}; currently running: {1}".format(
str(self.__scheduler), str(self.__scheduler.running)))
self.__scheduler.start()
def stop(self):
if not self.__scheduler is None:
self.__scheduler.shutdown(wait=False)
self.__scheduler = Scheduler(self.APS_CONFIG)
logger.info("[AlertScheduler] Stopped the alert scheduler.")
def reschedule(self):
"""
Removes jobs that are scheduled where their UUID no longer is valid.
Schedules jobs where the definition UUID is not currently scheduled.
"""
jobs_scheduled = 0
jobs_removed = 0
definitions = self.__load_definitions()
scheduled_jobs = self.__scheduler.get_jobs()
# for every scheduled job, see if its UUID is still valid
for scheduled_job in scheduled_jobs:
uuid_valid = False
for definition in definitions:
definition_uuid = definition.get_uuid()
if scheduled_job.name == definition_uuid:
uuid_valid = True
break
# jobs without valid UUIDs should be unscheduled
if uuid_valid == False:
jobs_removed += 1
logger.info("[AlertScheduler] Unscheduling {0}".format(scheduled_job.name))
self._collector.remove_by_uuid(scheduled_job.name)
self.__scheduler.unschedule_job(scheduled_job)
# for every definition, determine if there is a scheduled job
for definition in definitions:
definition_scheduled = False
for scheduled_job in scheduled_jobs:
definition_uuid = definition.get_uuid()
if definition_uuid == scheduled_job.name:
definition_scheduled = True
break
# if no jobs are found with the definitions UUID, schedule it
if definition_scheduled == False:
jobs_scheduled += 1
self.schedule_definition(definition)
logger.info("[AlertScheduler] Reschedule Summary: {0} rescheduled, {1} unscheduled".format(
str(jobs_scheduled), str(jobs_removed)))
def reschedule_all(self):
"""
Removes jobs that are scheduled where their UUID no longer is valid.
Schedules jobs where the definition UUID is not currently scheduled.
"""
jobs_scheduled = 0
jobs_removed = 0
definitions = self.__load_definitions()
scheduled_jobs = self.__scheduler.get_jobs()
# unschedule all scheduled jobs
for scheduled_job in scheduled_jobs:
jobs_removed += 1
logger.info("[AlertScheduler] Unscheduling {0}".format(scheduled_job.name))
self._collector.remove_by_uuid(scheduled_job.name)
self.__scheduler.unschedule_job(scheduled_job)
# for every definition, schedule a job
for definition in definitions:
jobs_scheduled += 1
self.schedule_definition(definition)
logger.info("[AlertScheduler] Reschedule Summary: {0} rescheduled, {1} unscheduled".format(
str(jobs_scheduled), str(jobs_removed)))
def collector(self):
""" gets the collector for reporting to the server """
return self._collector
def __load_definitions(self):
"""
Loads all alert definitions from a file. All clusters are stored in
a single file.
:return:
"""
definitions = []
all_commands = None
alerts_definitions_path = os.path.join(self.cachedir, self.FILENAME)
try:
with open(alerts_definitions_path) as fp:
all_commands = json.load(fp)
except:
logger.warning('[AlertScheduler] {0} not found or invalid. No alerts will be scheduled until registration occurs.'.format(alerts_definitions_path))
return definitions
for command_json in all_commands:
clusterName = '' if not 'clusterName' in command_json else command_json['clusterName']
hostName = '' if not 'hostName' in command_json else command_json['hostName']
for definition in command_json['alertDefinitions']:
alert = self.__json_to_callable(clusterName, hostName, definition)
if alert is None:
continue
alert.set_helpers(self._collector, self._cluster_configuration)
definitions.append(alert)
return definitions
def __json_to_callable(self, clusterName, hostName, json_definition):
"""
converts the json that represents all aspects of a definition
and makes an object that extends BaseAlert that is used for individual
"""
alert = None
try:
source = json_definition['source']
source_type = source.get('type', '')
if logger.isEnabledFor(logging.DEBUG):
logger.debug("[AlertScheduler] Creating job type {0} with {1}".format(source_type, str(json_definition)))
if source_type == AlertSchedulerHandler.TYPE_METRIC:
alert = MetricAlert(json_definition, source, self.config)
elif source_type == AlertSchedulerHandler.TYPE_PORT:
alert = PortAlert(json_definition, source)
elif source_type == AlertSchedulerHandler.TYPE_SCRIPT:
source['stacks_directory'] = self.stacks_dir
source['common_services_directory'] = self.common_services_dir
source['host_scripts_directory'] = self.host_scripts_dir
alert = ScriptAlert(json_definition, source, self.config)
elif source_type == AlertSchedulerHandler.TYPE_WEB:
alert = WebAlert(json_definition, source, self.config)
elif source_type == AlertSchedulerHandler.TYPE_RECOVERY:
alert = RecoveryAlert(json_definition, source, self.recovery_manger)
if alert is not None:
alert.set_cluster(clusterName, hostName)
except Exception,exception:
logger.exception("[AlertScheduler] Unable to load an invalid alert definition. It will be skipped.")
return alert
self.pin = pin
self.period = period
self.duration = duration
# Call on() in a new thread.
sched.add_interval_job(self.__on, seconds=self.period)
def __on(self):
log('%s On' % self.pin)
off_time = datetime.now() + timedelta(0, self.duration, 0)
# Schedule off() to run, in another thread, after 'off_time'.
sched.add_date_job(self.__off, off_time)
def __off(self):
log('%s Off' % self.pin)
def main():
Pwm("ding ", 1, SECONDS_UP)
Pwm(" DONG", 3, SECONDS_UP)
sched.start()
while True:
pass
if __name__ == '__main__':
sched = Scheduler(daemon=True)
atexit.register(lambda: sched.shutdown(wait=False))
main()
class Scheduler(object):
schedulr = None
aps3 = True
def __init__(self):
#####
# ApScheduler version detection
try:
# APScheduler 3.x implementation
from apscheduler.schedulers.background import BackgroundScheduler
self.schedulr = BackgroundScheduler()
self.aps3 = True
except ImportError:
# APScheduler 2.x implementation
from apscheduler.scheduler import Scheduler
self.schedulr = Scheduler()
self.aps3 = False
def start(self):
return self.schedulr.start()
def get_job(self, name):
if self.aps3:
return self.schedulr.get_job(name)
else:
jobs = self.schedulr.get_jobs()
for job in jobs:
if job.name == name:
return job
return None
def add_job(self, func, trigger, args=None, kwargs=None, id=None, **trigger_args):
if self.aps3:
return self.schedulr.add_job(func, trigger, id=id, args=args, kwargs=kwargs, **trigger_args)
else:
if trigger is 'date':
run_date = trigger_args['run_date'] # by intention: to raise if not set!
del trigger_args['run_date']
return self.schedulr.add_date_job(func, run_date, name=id, args=args, kwargs=kwargs)
elif trigger is 'interval':
# only partially implemented!!
seconds = 0
minutes = 0
hours = 0
if 'seconds' in trigger_args:
seconds = trigger_args.get('seconds', 0)
del trigger_args['seconds']
if 'minutes' in trigger_args:
minutes = trigger_args.get('minutes', 0)
del trigger_args['minutes']
if 'hours' in trigger_args:
hours = trigger_args.get('hours', 0)
del trigger_args['hours']
return self.schedulr.add_interval_job(func, name=id,
hours=hours, minutes=minutes, seconds=seconds,
args=args, kwargs=kwargs)
elif trigger is 'cron':
# only partially implemented!!
second = 0
minute = 0
hour = 0
if 'second' in trigger_args:
second = trigger_args.get('second', 0)
del trigger_args['second']
if 'minute' in trigger_args:
minute = trigger_args.get('minute', 0)
del trigger_args['minute']
if 'hour' in trigger_args:
hour = trigger_args.get('hour', 0)
del trigger_args['hour']
return self.schedulr.add_cron_job(func, name=id, hour=hour, minute=minute, second=second)
else:
raise NotImplementedError
def shutdown(self):
return self.schedulr.shutdown()
# https://github.com/ralphwetzel/theonionbox/issues/19#issuecomment-263110953
def check_tz(self):
from tzlocal import get_localzone
try:
# APScheduler 3.x
from apscheduler.util import astimezone
except ImportError:
# https://github.com/ralphwetzel/theonionbox/issues/31
# APScheduler 2.x
# import six
from pytz import timezone, utc
from datetime import tzinfo
# copied here from apscheduler/util.py (version 3.4)
# copyright Alex Grönholm
# https://github.com/agronholm/apscheduler
def astimezone(obj):
"""
Interprets an object as a timezone.
:rtype: tzinfo
"""
# if isinstance(obj, six.string_types):
if isinstance(obj, (str, unicode)):
return timezone(obj)
if isinstance(obj, tzinfo):
if not hasattr(obj, 'localize') or not hasattr(obj, 'normalize'):
raise TypeError('Only timezones from the pytz library are supported')
if obj.zone == 'local':
raise ValueError(
'Unable to determine the name of the local timezone -- you must explicitly '
'specify the name of the local timezone. Please refrain from using timezones like '
'EST to prevent problems with daylight saving time. Instead, use a locale based '
'timezone name (such as Europe/Helsinki).')
return obj
if obj is not None:
raise TypeError('Expected tzinfo, got %s instead' % obj.__class__.__name__)
tz = get_localzone()
try:
res = astimezone(tz)
except ValueError as ve:
return False
return True
class PyFlowScheduler(object):
"""
This object schedules the submission of the tasks in an :class:`Flow`.
There are two types of errors that might occur during the execution of the jobs:
#. Python exceptions
#. Abinit Errors.
Python exceptions are easy to detect and are usually due to a bug in abinitio or random errors such as IOError.
The set of Abinit Errors is much much broader. It includes wrong input data, segmentation
faults, problems with the resource manager, etc. Abinitio tries to handle the most common cases
but there's still a lot of room for improvement.
Note, in particular, that `PyFlowScheduler` will shutdown automatically if
#. The number of python exceptions is > MAX_NUM_PYEXC
#. The number of Abinit Errors (i.e. the number of tasks whose status is S_ERROR) is > MAX_NUM_ERRORS
#. The number of jobs launched becomes greater than (SAFETY_RATIO * total_number_of_tasks).
#. The scheduler will send an email to the user (specified by mailto) every REMINDME_S seconds.
If the mail cannot be sent, it will shutdown automatically.
This check prevents the scheduler from being trapped in an infinite loop.
"""
# Configuration file.
YAML_FILE = "scheduler.yml"
USER_CONFIG_DIR = os.path.join(os.getenv("HOME"), ".abinit", "abipy")
DEBUG = 0
Error = PyFlowSchedulerError
def __init__(self, **kwargs):
"""
Args:
weeks: number of weeks to wait
days: number of days to wait
hours: number of hours to wait
minutes: number of minutes to wait
seconds: number of seconds to wait
verbose: (int) verbosity level
max_njobs_inque: Limit on the number of jobs that can be present in the queue
use_dynamic_manager: True if the :class:`TaskManager` must be re-initialized from
file before launching the jobs. Default: False
max_nlaunch: Maximum number of tasks launched by radpifire (default -1 i.e. no limit)
"""
# Options passed to the scheduler.
self.sched_options = AttrDict(
weeks=kwargs.pop("weeks", 0),
days=kwargs.pop("days", 0),
hours=kwargs.pop("hours", 0),
minutes=kwargs.pop("minutes", 0),
seconds=kwargs.pop("seconds", 0),
#start_date=kwargs.pop("start_date", None),
)
if all(not v for v in self.sched_options.values()):
raise self.Error("Wrong set of options passed to the scheduler.")
self.mailto = kwargs.pop("mailto", None)
self.verbose = int(kwargs.pop("verbose", 0))
self.use_dynamic_manager = kwargs.pop("use_dynamic_manager", False)
self.max_njobs_inqueue = kwargs.pop("max_njobs_inqueue", 200)
self.REMINDME_S = float(kwargs.pop("REMINDME_S", 4 * 24 * 3600))
self.MAX_NUM_PYEXCS = int(kwargs.pop("MAX_NUM_PYEXCS", 0))
self.MAX_NUM_ABIERRS = int(kwargs.pop("MAX_NUM_ABIERRS", 0))
self.SAFETY_RATIO = int(kwargs.pop("SAFETY_RATIO", 5))
#self.MAX_ETIME_S = kwargs.pop("MAX_ETIME_S", )
self.max_nlaunch = kwargs.pop("max_nlaunch", -1)
if kwargs:
raise self.Error("Unknown arguments %s" % kwargs)
if has_sched_v3:
from apscheduler.schedulers.blocking import BlockingScheduler
self.sched = BlockingScheduler()
else:
from apscheduler.scheduler import Scheduler
self.sched = Scheduler(standalone=True)
self.nlaunch = 0
self.num_reminders = 1
# Used to keep track of the exceptions raised while the scheduler is running
self.exceptions = collections.deque(maxlen=self.MAX_NUM_PYEXCS + 10)
# Used to push additional info during the execution.
self.history = collections.deque(maxlen=100)
@classmethod
def from_file(cls, filepath):
"""Read the configuration parameters from a Yaml file."""
with open(filepath, "r") as fh:
return cls(**yaml.load(fh))
@classmethod
def from_string(cls, s):
"""Create an istance from string s containing a YAML dictionary."""
stream = cStringIO(s)
stream.seek(0)
return cls(**yaml.load(stream))
@classmethod
def from_user_config(cls):
"""
Initialize the :class:`PyFlowScheduler` from the YAML file 'scheduler.yml'.
Search first in the working directory and then in the configuration directory of abipy.
Raises:
RuntimeError if file is not found.
"""
# Try in the current directory.
path = os.path.join(os.getcwd(), cls.YAML_FILE)
if os.path.exists(path):
return cls.from_file(path)
# Try in the configuration directory.
path = os.path.join(cls.USER_CONFIG_DIR, cls.YAML_FILE)
if os.path.exists(path):
return cls.from_file(path)
err_msg = "Cannot locate %s neither in current directory nor in %s" % (
cls.YAML_FILE, path)
raise cls.Error(err_msg)
def __str__(self):
"""String representation."""
lines = [self.__class__.__name__ + ", Pid: %d" % self.pid]
app = lines.append
app("Scheduler options: %s" % str(self.sched_options))
app(80 * "=")
app(str(self.flow))
return "\n".join(lines)
@property
def pid(self):
"""The pid of the process associated to the scheduler."""
try:
return self._pid
except AttributeError:
self._pid = os.getpid()
return self._pid
@property
def pid_file(self):
"""
Absolute path of the file with the pid.
The file is located in the workdir of the flow
"""
return self._pid_file
@property
def flow(self):
"""`Flow`."""
return self._flow
@property
def num_excs(self):
"""Number of exceptions raised so far."""
return len(self.exceptions)
def get_delta_etime(self):
"""Returns a `timedelta` object representing with the elapsed time."""
return timedelta(seconds=(time.time() - self.start_time))
def add_flow(self, flow):
"""Add an :class:`Flow` flow to the scheduler."""
if hasattr(self, "_flow"):
raise self.Error("Only one flow can be added to the scheduler.")
pid_file = os.path.join(flow.workdir, "_PyFlowScheduler.pid")
if os.path.isfile(pid_file):
flow.show_status()
err_msg = ("""
pid_file %s already exists
There are two possibilities:
1) There's an another instance of PyFlowScheduler running
2) The previous scheduler didn't exit in a clean way
To solve case 1:
Kill the previous scheduler (use 'kill pid' where pid is the number reported in the file)
Then you can restart the new scheduler.
To solve case 2:
Remove the pid_file and restart the scheduler.
Exiting""" % pid_file)
raise self.Error(err_msg)
with open(pid_file, "w") as fh:
fh.write(str(self.pid))
self._pid_file = pid_file
self._flow = flow
def start(self):
"""
Starts the scheduler in a new thread. Returns True if success.
In standalone mode, this method will block until there are no more scheduled jobs.
"""
self.history.append("Started on %s" % time.asctime())
self.start_time = time.time()
if has_sched_v3:
self.sched.add_job(self.callback, "interval", **self.sched_options)
else:
self.sched.add_interval_job(self.callback, **self.sched_options)
errors = self.flow.look_before_you_leap()
if errors:
print(errors)
self.exceptions.append(errors)
return False
# Try to run the job immediately. If something goes wrong return without initializing the scheduler.
self._runem_all()
if self.exceptions:
self.cleanup()
self.send_email(
msg=
"Error while trying to run the flow for the first time!\n %s" %
self.exceptions)
return False
self.sched.start()
return True
def _runem_all(self):
"""
This function checks the status of all tasks,
tries to fix tasks that went unconverged, abicritical, or queuecritical
and tries to run all the tasks that can be submitted.+
"""
excs = []
flow = self.flow
# Allow to change the manager at run-time
if self.use_dynamic_manager:
from pymatgen.io.abinitio.tasks import TaskManager
new_manager = TaskManager.from_user_config()
for work in flow:
work.set_manager(new_manager)
nqjobs = flow.get_njobs_in_queue()
if nqjobs is None:
nqjobs = 0
print('Cannot get njobs_inqueue')
if nqjobs >= self.max_njobs_inqueue:
print("Too many jobs in the queue, returning")
return
if self.max_nlaunch == -1:
max_nlaunch = self.max_njobs_inqueue - nqjobs
else:
max_nlaunch = min(self.max_njobs_inqueue - nqjobs,
self.max_nlaunch)
# check status
flow.check_status()
flow.show_status()
# fix problems
# Try to restart the unconverged tasks
# todo donot fire here but prepare for fireing in rapidfire
for task in self.flow.unconverged_tasks:
try:
logger.info("Flow will try restart task %s" % task)
fired = task.restart()
if fired:
self.nlaunch += 1
max_nlaunch -= 1
if max_nlaunch == 0:
print("Restart: too many jobs in the queue, returning")
flow.pickle_dump()
return
except Exception:
excs.append(straceback())
# move here from withing rapid fire ...
# fix only prepares for restarting, and sets to ready
flow.fix_critical()
# update database
flow.pickle_dump()
#if self.num_restarts == self.max_num_restarts:
# info_msg = "Reached maximum number of restarts. Cannot restart anymore Returning"
# logger.info(info_msg)
# self.history.append(info_msg)
# return 1
# Submit the tasks that are ready.
try:
nlaunch = PyLauncher(flow).rapidfire(max_nlaunch=max_nlaunch,
sleep_time=10)
self.nlaunch += nlaunch
if nlaunch:
print("[%s] Number of launches: %d" %
(time.asctime(), nlaunch))
except Exception:
excs.append(straceback())
flow.show_status()
if excs:
logger.critical("*** Scheduler exceptions:\n *** %s" %
"\n".join(excs))
self.exceptions.extend(excs)
def callback(self):
"""The function that will be executed by the scheduler."""
try:
return self._callback()
except:
# All exceptions raised here will trigger the shutdown!
self.exceptions.append(straceback())
self.shutdown(msg="Exception raised in callback!")
def _callback(self):
"""The actual callback."""
if self.DEBUG:
# Show the number of open file descriptors
print(">>>>> _callback: Number of open file descriptors: %s" %
get_open_fds())
#print('before _runem_all in _callback')
self._runem_all()
# Mission accomplished. Shutdown the scheduler.
all_ok = self.flow.all_ok
if self.verbose:
print("all_ok", all_ok)
if all_ok:
self.shutdown(
msg=
"All tasks have reached S_OK. Will shutdown the scheduler and exit"
)
# Handle failures.
err_msg = ""
# Shall we send a reminder to the user?
delta_etime = self.get_delta_etime()
if delta_etime.total_seconds() > self.num_reminders * self.REMINDME_S:
self.num_reminders += 1
msg = (
"Just to remind you that the scheduler with pid %s, flow %s\n has been running for %s "
% (self.pid, self.flow, delta_etime))
retcode = self.send_email(msg, tag="[REMINDER]")
if retcode:
# Cannot send mail, shutdown now!
msg += (
"\nThe scheduler tried to send an e-mail to remind the user\n"
+ " but send_email returned %d. Aborting now" % retcode)
err_msg += msg
#if delta_etime.total_seconds() > self.MAX_ETIME_S:
# err_msg += "\nExceeded MAX_ETIME_S %s. Will shutdown the scheduler and exit" % self.MAX_ETIME_S
# Too many exceptions. Shutdown the scheduler.
if self.num_excs > self.MAX_NUM_PYEXCS:
msg = "Number of exceptions %s > %s. Will shutdown the scheduler and exit" % (
self.num_excs, self.MAX_NUM_PYEXCS)
err_msg += boxed(msg)
# Paranoid check: disable the scheduler if we have submitted
# too many jobs (it might be due to some bug or other external reasons
# such as race conditions between difference callbacks!)
if self.nlaunch > self.SAFETY_RATIO * self.flow.num_tasks:
msg = "Too many jobs launched %d. Total number of tasks = %s, Will shutdown the scheduler and exit" % (
self.nlaunch, self.flow.num_tasks)
err_msg += boxed(msg)
# Count the number of tasks with status == S_ERROR.
if self.flow.num_errored_tasks > self.MAX_NUM_ABIERRS:
msg = "Number of tasks with ERROR status %s > %s. Will shutdown the scheduler and exit" % (
self.flow.num_errored_tasks, self.MAX_NUM_ABIERRS)
err_msg += boxed(msg)
# Count the number of tasks with status == S_UNCONVERGED.
#if self.flow.num_unconverged_tasks:
# # TODO: this is needed to avoid deadlocks, automatic restarting is not available yet
# msg = ("Found %d unconverged tasks."
# "Automatic restarting is not available yet. Will shutdown the scheduler and exit"
# % self.flow.num_unconverged_tasks)
# err_msg += boxed(msg)
#deadlocks = self.detect_deadlocks()
#if deadlocks:
# msg = ("Detected deadlocks in flow. Will shutdown the scheduler and exit"
# % self.flow.num_unconverged_tasks)
# err_msg += boxed(msg)
if err_msg:
# Something wrong. Quit
self.shutdown(err_msg)
return len(self.exceptions)
def cleanup(self):
"""Cleanup routine: remove the pid file and save the pickle database"""
try:
os.remove(self.pid_file)
except OSError:
logger.critical("Could not remove pid_file")
pass
# Save the final status of the flow.
self.flow.pickle_dump()
def shutdown(self, msg):
"""Shutdown the scheduler."""
try:
self.cleanup()
#if False and self.flow.has_db:
# try:
# self.flow.db_insert()
# except Exception:
# logger.critical("MongoDb insertion failed.")
self.history.append("Completed on %s" % time.asctime())
self.history.append("Elapsed time %s" % self.get_delta_etime())
if self.DEBUG:
print(">>>>> shutdown: Number of open file descriptors: %s" %
get_open_fds())
retcode = self.send_email(msg)
if self.DEBUG:
print("send_mail retcode", retcode)
# Write file with the list of exceptions:
if self.exceptions:
dump_file = os.path.join(self.flow.workdir, "_exceptions")
with open(dump_file, "w") as fh:
fh.writelines(self.exceptions)
fh.write("Shutdown message:\n%s" % msg)
finally:
# Shutdown the scheduler thus allowing the process to exit.
print('this should be the shutdown of the scheduler')
# Unschedule all the jobs before calling shutdown
self.sched.print_jobs()
for job in self.sched.get_jobs():
self.sched.unschedule_job(job)
self.sched.print_jobs()
self.sched.shutdown()
# Uncomment the line below if shutdown does not work!
#os.system("kill -9 %d" % os.getpid())
def send_email(self, msg, tag=None):
"""
Send an e-mail before completing the shutdown.
Returns 0 if success.
"""
try:
return self._send_email(msg, tag)
except:
self.exceptions.append(straceback())
return -2
def _send_email(self, msg, tag):
if self.mailto is None:
return -1
header = msg.splitlines()
app = header.append
app("Submitted on %s" % time.ctime(self.start_time))
app("Completed on %s" % time.asctime())
app("Elapsed time %s" % str(self.get_delta_etime()))
app("Number of errored tasks: %d" % self.flow.num_errored_tasks)
app("Number of unconverged tasks: %d" %
self.flow.num_unconverged_tasks)
strio = cStringIO()
strio.writelines("\n".join(header) + 4 * "\n")
# Add the status of the flow.
self.flow.show_status(stream=strio)
if self.exceptions:
# Report the list of exceptions.
strio.writelines(self.exceptions)
if tag is None:
tag = " [ALL OK]" if self.flow.all_ok else " [WARNING]"
return sendmail(subject=self.flow.name + tag,
text=strio.getvalue(),
mailto=self.mailto)
class AlertSchedulerHandler():
FILENAME = 'definitions.json'
TYPE_PORT = 'PORT'
TYPE_METRIC = 'METRIC'
TYPE_SCRIPT = 'SCRIPT'
TYPE_WEB = 'WEB'
APS_CONFIG = {
'threadpool.core_threads': 3,
'coalesce': True,
'standalone': False
}
def __init__(self, cachedir, stacks_dir, common_services_dir, host_scripts_dir,
cluster_configuration, config, in_minutes=True):
self.cachedir = cachedir
self.stacks_dir = stacks_dir
self.common_services_dir = common_services_dir
self.host_scripts_dir = host_scripts_dir
self._cluster_configuration = cluster_configuration
if not os.path.exists(cachedir):
try:
os.makedirs(cachedir)
except:
logger.critical("[AlertScheduler] Could not create the cache directory {0}".format(cachedir))
self._collector = AlertCollector()
self.__scheduler = Scheduler(AlertSchedulerHandler.APS_CONFIG)
self.__in_minutes = in_minutes
self.config = config
# register python exit handler
atexit.register(self.exit_handler)
def exit_handler(self):
"""
Exit handler
"""
self.stop()
def update_definitions(self, heartbeat):
"""
Updates the persisted alert definitions JSON.
:param heartbeat:
:return:
"""
if 'alertDefinitionCommands' not in heartbeat:
logger.warning("There are no alert definition commands in the heartbeat; unable to update definitions")
return
# prune out things we don't want to store
alert_definitions = []
for command in heartbeat['alertDefinitionCommands']:
command_copy = command.copy()
# no need to store these since we always use the in-memory cached values
if 'configurations' in command_copy:
del command_copy['configurations']
alert_definitions.append(command_copy)
# write out the new definitions
with open(os.path.join(self.cachedir, self.FILENAME), 'w') as f:
json.dump(alert_definitions, f, indent=2)
# reschedule only the jobs that have changed
self.reschedule()
def __make_function(self, alert_def):
return lambda: alert_def.collect()
def start(self):
""" loads definitions from file and starts the scheduler """
if self.__scheduler is None:
return
if self.__scheduler.running:
self.__scheduler.shutdown(wait=False)
self.__scheduler = Scheduler(AlertSchedulerHandler.APS_CONFIG)
alert_callables = self.__load_definitions()
# schedule each definition
for _callable in alert_callables:
self.schedule_definition(_callable)
logger.info("[AlertScheduler] Starting {0}; currently running: {1}".format(
str(self.__scheduler), str(self.__scheduler.running)))
self.__scheduler.start()
def stop(self):
if not self.__scheduler is None:
self.__scheduler.shutdown(wait=False)
self.__scheduler = Scheduler(AlertSchedulerHandler.APS_CONFIG)
logger.info("[AlertScheduler] Stopped the alert scheduler.")
def reschedule(self):
"""
Removes jobs that are scheduled where their UUID no longer is valid.
Schedules jobs where the definition UUID is not currently scheduled.
"""
jobs_scheduled = 0
jobs_removed = 0
definitions = self.__load_definitions()
scheduled_jobs = self.__scheduler.get_jobs()
# for every scheduled job, see if its UUID is still valid
for scheduled_job in scheduled_jobs:
uuid_valid = False
for definition in definitions:
definition_uuid = definition.get_uuid()
if scheduled_job.name == definition_uuid:
uuid_valid = True
break
# jobs without valid UUIDs should be unscheduled
if uuid_valid == False:
jobs_removed += 1
logger.info("[AlertScheduler] Unscheduling {0}".format(scheduled_job.name))
self._collector.remove_by_uuid(scheduled_job.name)
self.__scheduler.unschedule_job(scheduled_job)
# for every definition, determine if there is a scheduled job
for definition in definitions:
definition_scheduled = False
for scheduled_job in scheduled_jobs:
definition_uuid = definition.get_uuid()
if definition_uuid == scheduled_job.name:
definition_scheduled = True
break
# if no jobs are found with the definitions UUID, schedule it
if definition_scheduled == False:
jobs_scheduled += 1
self.schedule_definition(definition)
logger.info("[AlertScheduler] Reschedule Summary: {0} rescheduled, {1} unscheduled".format(
str(jobs_scheduled), str(jobs_removed)))
def reschedule_all(self):
"""
Removes jobs that are scheduled where their UUID no longer is valid.
Schedules jobs where the definition UUID is not currently scheduled.
"""
jobs_scheduled = 0
jobs_removed = 0
definitions = self.__load_definitions()
scheduled_jobs = self.__scheduler.get_jobs()
# unschedule all scheduled jobs
for scheduled_job in scheduled_jobs:
jobs_removed += 1
logger.info("[AlertScheduler] Unscheduling {0}".format(scheduled_job.name))
self._collector.remove_by_uuid(scheduled_job.name)
self.__scheduler.unschedule_job(scheduled_job)
# for every definition, schedule a job
for definition in definitions:
jobs_scheduled += 1
self.schedule_definition(definition)
logger.info("[AlertScheduler] Reschedule Summary: {0} rescheduled, {1} unscheduled".format(
str(jobs_scheduled), str(jobs_removed)))
def collector(self):
""" gets the collector for reporting to the server """
return self._collector
def __load_definitions(self):
"""
Loads all alert definitions from a file. All clusters are stored in
a single file.
:return:
"""
definitions = []
all_commands = None
alerts_definitions_path = os.path.join(self.cachedir, self.FILENAME)
try:
with open(alerts_definitions_path) as fp:
all_commands = json.load(fp)
except:
logger.warning('[AlertScheduler] {0} not found or invalid. No alerts will be scheduled until registration occurs.'.format(alerts_definitions_path))
return definitions
for command_json in all_commands:
clusterName = '' if not 'clusterName' in command_json else command_json['clusterName']
hostName = '' if not 'hostName' in command_json else command_json['hostName']
for definition in command_json['alertDefinitions']:
alert = self.__json_to_callable(clusterName, hostName, definition)
if alert is None:
continue
alert.set_helpers(self._collector, self._cluster_configuration)
definitions.append(alert)
return definitions
def __json_to_callable(self, clusterName, hostName, json_definition):
"""
converts the json that represents all aspects of a definition
and makes an object that extends BaseAlert that is used for individual
"""
source = json_definition['source']
source_type = source.get('type', '')
if logger.isEnabledFor(logging.DEBUG):
logger.debug("[AlertScheduler] Creating job type {0} with {1}".format(source_type, str(json_definition)))
alert = None
if source_type == AlertSchedulerHandler.TYPE_METRIC:
alert = MetricAlert(json_definition, source)
elif source_type == AlertSchedulerHandler.TYPE_PORT:
alert = PortAlert(json_definition, source)
elif source_type == AlertSchedulerHandler.TYPE_SCRIPT:
source['stacks_directory'] = self.stacks_dir
source['common_services_directory'] = self.common_services_dir
source['host_scripts_directory'] = self.host_scripts_dir
alert = ScriptAlert(json_definition, source, self.config)
elif source_type == AlertSchedulerHandler.TYPE_WEB:
alert = WebAlert(json_definition, source, self.config)
if alert is not None:
alert.set_cluster(clusterName, hostName)
return alert
def schedule_definition(self,definition):
"""
Schedule a definition (callable). Scheduled jobs are given the UUID
as their name so that they can be identified later on.
<p/>
This function can be called with a definition that is disabled; it will
simply NOOP.
"""
# NOOP if the definition is disabled; don't schedule it
if not definition.is_enabled():
logger.info("[AlertScheduler] The alert {0} with UUID {1} is disabled and will not be scheduled".format(
definition.get_name(),definition.get_uuid()))
return
job = None
if self.__in_minutes:
job = self.__scheduler.add_interval_job(self.__make_function(definition),
minutes=definition.interval())
else:
job = self.__scheduler.add_interval_job(self.__make_function(definition),
seconds=definition.interval())
# although the documentation states that Job(kwargs) takes a name
# key/value pair, it does not actually set the name; do it manually
if job is not None:
job.name = definition.get_uuid()
logger.info("[AlertScheduler] Scheduling {0} with UUID {1}".format(
definition.get_name(), definition.get_uuid()))
def get_job_count(self):
"""
Gets the number of jobs currently scheduled. This is mainly used for
test verification of scheduling.
"""
if self.__scheduler is None:
return 0
return len(self.__scheduler.get_jobs())
def execute_alert(self, execution_commands):
"""
Executes an alert immediately, ignoring any scheduled jobs. The existing
jobs remain untouched. The result of this is stored in the alert
collector for tranmission during the next heartbeat
"""
if self.__scheduler is None or execution_commands is None:
return
for execution_command in execution_commands:
try:
alert_definition = execution_command['alertDefinition']
clusterName = '' if not 'clusterName' in execution_command else execution_command['clusterName']
hostName = '' if not 'hostName' in execution_command else execution_command['hostName']
alert = self.__json_to_callable(clusterName, hostName, alert_definition)
if alert is None:
continue
logger.info("[AlertScheduler] Executing on-demand alert {0} ({1})".format(alert.get_name(),
alert.get_uuid()))
alert.set_helpers(self._collector, self._cluster_configuration)
alert.collect()
except:
logger.exception("[AlertScheduler] Unable to execute the alert outside of the job scheduler")
class TestOfflineScheduler(object):
def setup(self):
self.scheduler = Scheduler()
def teardown(self):
if self.scheduler.running:
self.scheduler.shutdown()
@raises(KeyError)
def test_jobstore_twice(self):
self.scheduler.add_jobstore(RAMJobStore(), 'dummy')
self.scheduler.add_jobstore(RAMJobStore(), 'dummy')
def test_add_tentative_job(self):
job = self.scheduler.add_date_job(lambda: None,
datetime(2200, 7, 24),
jobstore='dummy')
assert isinstance(job, Job)
eq_(self.scheduler.get_jobs(), [])
def test_configure_jobstore(self):
conf = {
'apscheduler.jobstore.ramstore.class':
'apscheduler.jobstores.ram_store:RAMJobStore'
}
self.scheduler.configure(conf)
self.scheduler.remove_jobstore('ramstore')
def test_shutdown_offline(self):
self.scheduler.shutdown()
def test_configure_no_prefix(self):
global_options = {'misfire_grace_time': '2', 'daemonic': 'false'}
self.scheduler.configure(global_options)
eq_(self.scheduler.misfire_grace_time, 1)
eq_(self.scheduler.daemonic, True)
def test_configure_prefix(self):
global_options = {
'apscheduler.misfire_grace_time': 2,
'apscheduler.daemonic': False
}
self.scheduler.configure(global_options)
eq_(self.scheduler.misfire_grace_time, 2)
eq_(self.scheduler.daemonic, False)
def test_add_listener(self):
val = []
self.scheduler.add_listener(val.append)
event = SchedulerEvent(EVENT_SCHEDULER_START)
self.scheduler._notify_listeners(event)
eq_(len(val), 1)
eq_(val[0], event)
event = SchedulerEvent(EVENT_SCHEDULER_SHUTDOWN)
self.scheduler._notify_listeners(event)
eq_(len(val), 2)
eq_(val[1], event)
self.scheduler.remove_listener(val.append)
self.scheduler._notify_listeners(event)
eq_(len(val), 2)
def test_pending_jobs(self):
# Tests that pending jobs are properly added to the jobs list when
# the scheduler is started (and not before!)
self.scheduler.add_date_job(lambda: None, datetime(9999, 9, 9))
eq_(self.scheduler.get_jobs(), [])
self.scheduler.start()
jobs = self.scheduler.get_jobs()
eq_(len(jobs), 1)
def tag_check_task(tag, min=1):
cron = Scheduler(daemon=True)
cron.start()
@cron.interval_schedule(minutes=min)
def job_function():
now_day = datetime.now().date()
tag_db = Tags().get_or_create(tag, now_day)
tag_list = tag_filter(tag, 100)
tag_list.reverse()
new_tag = {}
for tags in tag_list:
_create_time = datetime.strptime(
tags['created'], '%Y-%m-%dT%H:%M:%S') + timedelta(hours=3)
if now_day.day == _create_time.day:
new_tag[tags['id']] = tags
post_list = list(new_tag.keys())
post_list.sort()
if not tag_db.last:
tag_db.last = post_list[-1]
post_ids = tag_db.last
db.session.add(tag_db)
db.session.commit()
else:
try:
post_list = post_list[post_list.index(tag_db.last) + 1:]
except:
last_post = Posts().query.order_by('-id').first()
tag_db.last = None
db.session.delete(last_post)
db.session.add(tag_db)
db.session.commit()
last_tag_get_update()
return
if not len(post_list):
return
for post_db in post_list:
if not Posts.query.filter_by(post_id=post_db,
tag=tag_db).first() or False:
_post = Posts()
_post.post_id = post_db
_post.author = new_tag[post_db]['author']
_post.author = new_tag[post_db]['author']
_post.title = new_tag[post_db]['title']
_post.url = new_tag[post_db]['url']
_post.tag = tag_db
_post.date = datetime.strptime(
new_tag[post_db]['created'],
'%Y-%m-%dT%H:%M:%S') + timedelta(hours=3)
db.session.add(_post)
db.session.commit()
tag_db.last = post_db
db.session.add(tag_db)
db.session.commit()
atexit.register(lambda: cron.shutdown(wait=False))
class PeetsMediaTranslator(DatagramProtocol):
''' A translator protocol to relay local udp traffic to NDN and remote NDN traffic to local udp.
This class also implements the strategy for fetching remote data.
If the remote seq is unknown, use a short prefix without seq to probe;
otherwise use a naive leaking-bucket like method to fetch the remote data
We seperate the fetching of the media stream and the fetching of the control stream (RTCP, STUN, etc).
'''
__logger = Logger.get_logger('PeetsMediaTranslator')
def __init__(self, factory, pipe_size):
'''
Args:
factory (PeetsServerFactory) : the factory that stores necessory information about the local user
pipe_size (int) : the pipeline size for fetching the remote media stream. Pipelining allows us to minimize impact of the interest-data roundtrip delay.
'''
self.factory = factory
self.pipe_size = pipe_size
self.factory = factory
self.factory.set_local_status_callback(self.toggle_scheduler)
# here we use two sockets, because the pending interests sent by a socket can not be satisified
# by the content published later by the same socket
self.ccnx_int_socket = CcnxSocket()
self.ccnx_int_socket.start()
self.ccnx_con_socket = CcnxSocket()
self.ccnx_con_socket.start()
self.stream_closure = PeetsClosure(msg_callback = self.stream_callback, timeout_callback = self.stream_timeout_callback)
self.probe_closure = PeetsClosure(msg_callback = self.probe_callback, timeout_callback = self.probe_timeout_callback)
self.ctrl_probe_closure = PeetsClosure(msg_callback = self.ctrl_probe_callback, timeout_callback = self.ctrl_probe_timeout_callback)
self.scheduler = None
self.peets_status = None
def toggle_scheduler(self, status):
'''Start or stop the scheduler for periodic jobs.
Args:
status (str): either 'Running' or 'Stopped'
'''
if status == 'Running':
self.peets_status = 'Running'
self.scheduler = Scheduler()
self.scheduler.start()
self.scheduler.add_interval_job(self.fetch_media, seconds = 0.01, max_instances = 2)
elif status == 'Stopped':
self.peets_status = 'Stopped'
for job in self.scheduler.get_jobs():
self.scheduler.unschedule_job(job)
self.scheduler.shutdown(wait = True)
self.scheduler = None
def datagramReceived(self, data, (host, port)):
'''Intercept the webrtc traffice from the local front end and relay it to the NDN
Args:
data (bytes) : the UDP data
host (str) : the IP of the source
port (int) : the port of the source
1. Differentiate RTP vs RTCP
RTCP: packet type (PT) = 200 - 208
SR (sender report) 200
RR (receiver report) 201
SDES (source description) 202
BYE (goodbye) 203
App (application-defined) 204
other types go until 208
RFC 5761 (implemented by WebRTC) makes sure that RTP's PT field
plus M field (which is equal to the PT field in RTCP) would not conflict
2. Differentiate STUN vs RTP & RTCP
STUN: the most significant 2 bits of every STUN msg MUST be zeros (RFC 5389)
RTP & RTCP: version bits (2 bits) value equals 2
Note:
Tried to fake a Stun request and response so that we don't have to relay stun msgs to NDN, but failed. It worked for a time, although will significantly high rate of the STUN message exchanges
We need to use the username exchanged in the sdps for stun it worked for a while but magically stopped working, so now we still send it over NDN
Note 2:
We only publish one medai stream from the local user (with the default offer SDP). We publish RTCP and STUN for each PeerConnections though.
'''
# mask to test most significant 2 bits
msg = bytearray(data)
c = self.factory.client
if msg[0] & 0xC0 == 0 or msg[1] > 199 and msg[1] < 209:
try:
ctrl_seq = c.ctrl_seqs[port]
cid = c.remote_cids[port]
# RTCP and STUN is for each peerconnection. the cid of remote user is used to identify the peer connection so that remote user knows which one to fetch
name = c.local_user.get_ctrl_prefix() + '/' + cid + '/' + str(ctrl_seq)
c.ctrl_seqs[port] = ctrl_seq + 1
self.ccnx_con_socket.publish_content(name, data)
except KeyError:
pass
elif c.media_source_port == port:
# only publish one media stream
name = c.local_user.get_media_prefix() + '/' + str(c.local_seq)
c.local_seq += 1
self.ccnx_con_socket.publish_content(name, data)
class Subject:
def __init__(self):
self.gender = 0
self.mood = []
self.scheduler = None
self.name = "(no collaborator)"
def evaluate(self):
current = int(time.time())
timeView = list(
filter(lambda element: current - CONST_TIME <= element[0],
self.mood))
happyView = list(
filter(lambda element: element[1] == "happy", timeView))
if (len(timeView) == 0):
self.unload()
elif len(happyView) / len(timeView) * 100 < CONST_TREESHOLD:
self.unload()
def unload(self):
urllib.request.urlopen(CONST_BUZZ).read()
def start(self):
self.scheduler = Scheduler()
self.scheduler.start()
self.scheduler.add_interval_job(self.evaluate, seconds=CONST_TIME)
def stop(self):
self.scheduler.shutdown()
def add(self, labels, mood, gender):
labeledMoods = np.array(list(zip(labels, mood)))
timestamp = int(time.time())
self.gender = gender
for labeledMood in labeledMoods:
self.mood.append((timestamp, labeledMood[0], labeledMood[1]))
def persist(self, mood, gender, timestamp):
mongo = MongoClient(MONGO_CONNECTION)
database = mongo[MONGO_DATABASE]
post = {
"collaborator": self.name,
"mood": mood,
"gender": gender,
"timestamp": timestamp
}
database[COLLECTION].insert_one(post)
def addMood(self, mood, gender):
timestamp = int(time.time())
self.gender = gender
self.mood.append((timestamp, mood, 1))
self.persist(mood, gender, timestamp)
__author__ = 'wenychan'
# For version apscheduler 2.1.2
import time
import atexit
from apscheduler.scheduler import Scheduler
# Start the scheduler
sched = Scheduler(daemonic=True)
# register three exit handlers
atexit.register(lambda: sched.shutdown())
sched.start()
def job_function():
print "Hello World"
time.sleep(3)
print "exit Hello World"
# Schedule job_function to be called every two hours
sched.add_interval_job(job_function, seconds=1, name='test_job1', max_instances=1)
sched.add_interval_job(job_function, seconds=3, name='test_job2', max_instances=1, max_runs=1) # sched.add_interval_job(job_function, seconds=5, name='test_job3', max_instances=1)
sched.add_interval_job(job_function, seconds=5, name='test_job3', max_instances=1)
sched.add_interval_job(job_function, seconds=6, name='test_job4', max_instances=1)
sched.add_interval_job(job_function, seconds=3, name='test_job5', max_instances=1)
sched.add_interval_job(job_function, seconds=1, name='test_job6', max_instances=1)
print len(sched.get_jobs())
time.sleep(5)
print len(sched.get_jobs())
time.sleep(50)
def init_scheduler():
scheduler = Scheduler()
atexit.register(lambda: scheduler.shutdown(wait=False))
return scheduler
mC(auth['account_sid'],auth['auth_token'],targetNumber,ngrokURL)
print "Call Queued Up"
time.sleep(60) # Twilio default call timeout is 60 sec
#kill = raw_input("Kill proceses? (y/n):")
#while kill in ['n']:
# kill = raw_input("Kill proceses? (y/n):")
print "Killing processes"
ngrok.kill() # kill ngrok process
os.system('killall -KILL Python') # kill web server and this thread
# Start the scheduler
sched = Scheduler()
sched.start()
# Convert UTC target to local time
localTarget = utc2local(datetime.strptime(targetTimeUTC,'%Y-%m-%d %H:%M:%S'))
job = sched.add_date_job(main, localTarget, [filename,targetNumber,ngrokURL])
sched.print_jobs()
print 'Current time is %s' % datetime.now()
# Keep scheduler alive until you hit Ctrl+C!
while True:
time.sleep(1)
sched.shutdown()
class LocalScheduler(object):
scheduler_registry = {}
_lockdown = False
@classmethod
def get(cls, name):
return cls.scheduler_registry[name]
@classmethod
def get_all(cls):
return cls.scheduler_registry.values()
@classmethod
def shutdown_all(cls):
for scheduler in cls.scheduler_registry.values():
scheduler.stop()
@classmethod
def lockdown(cls):
cls._lockdown = True
@classmethod
def clear_all(cls):
for scheduler in cls.scheduler_registry.values():
scheduler.clear()
def __init__(self, name, label=None):
self.scheduled_jobs = {}
self._scheduler = None
self.name = name
self.label = label
self.__class__.scheduler_registry[self.name] = self
def start(self):
logger.info('Starting scheduler: %s' % self.name)
if not self.__class__._lockdown:
self._scheduler = OriginalScheduler()
for job in self.scheduled_jobs.values():
self._schedule_job(job)
self._scheduler.start()
else:
logger.debug('lockdown in effect')
def stop(self):
if self._scheduler:
self._scheduler.shutdown()
del self._scheduler
self._scheduler = None
@property
def running(self):
if self._scheduler:
return self._scheduler.running
else:
return False
def clear(self):
for job in self.scheduled_jobs.values():
self.stop_job(job)
def stop_job(self, job):
if self.running:
self._scheduler.unschedule_job(job._job)
del (self.scheduled_jobs[job.name])
job.scheduler = None
def _schedule_job(self, job):
if isinstance(job, IntervalJob):
job._job = self._scheduler.add_interval_job(
job.function, *job.args, **job.kwargs)
elif isinstance(job, DateJob):
job._job = self._scheduler.add_date_job(job.function, *job.args,
**job.kwargs)
elif isinstance(job, CronJob):
job._job = self._scheduler.add_cron_job(job.function, *job.args,
**job.kwargs)
else:
raise UnknownJobClass
def add_job(self, job):
logger.debug('adding job')
if job.scheduler or job.name in self.scheduled_jobs.keys():
raise AlreadyScheduled
if self._scheduler:
self._schedule_job(job)
job.scheduler = self
self.scheduled_jobs[job.name] = job
def add_interval_job(self, name, label, function, *args, **kwargs):
job = IntervalJob(name=name,
label=label,
function=function,
*args,
**kwargs)
self.add_job(job)
return job
def add_date_job(self, name, label, function, *args, **kwargs):
job = DateJob(name=name,
label=label,
function=function,
*args,
**kwargs)
self.add_job(job)
return job
def add_cron_job(self, name, label, function, *args, **kwargs):
job = CronJob(name=name,
label=label,
function=function,
*args,
**kwargs)
self.add_job(job)
return job
def get_job_list(self):
return self.scheduled_jobs.values()
def get_job_by_name(self, name):
try:
return self.scheduled_jobs[name]
except KeyError:
raise UnknownJob
def __unicode__(self):
return unicode(self.label or self.name)
end_time = datetime.strptime(event['event_end_time'], time_format)
event_duration = (end_time - start_time).seconds
sched.add_cron_job(led_chain.auto_transition,
hour=start_hour,
minute=start_minute,
second=start_second,
name=event['event_name'],
kwargs={
'state': event['event_state'],
'transition_duration':
event['transition_duration']
},
misfire_grace_time=event_duration)
app.logger.debug("Startup job list contains : %s" % sched.get_jobs())
try:
app.run(host='0.0.0.0',
port=int(app_config.get("general", "web_port")),
use_reloader=False)
except KeyboardInterrupt:
app.logger.warning("Caught keyboard interupt. Shutting down ...")
app.logger.info("Calling shutdown on led chain")
led_chain.shutdown()
app.logger.info("Calling shutdown on scheduler")
sched.shutdown(wait=False)
app.logger.info("Shutting down logger and exiting ...")
logging.shutdown()
exit(0)
class PyFlowScheduler(object):
"""
This object schedules the submission of the tasks in a :class:`Flow`.
There are two types of errors that might occur during the execution of the jobs:
#. Python exceptions
#. Errors in the ab-initio code
Python exceptions are easy to detect and are usually due to a bug in the python code or random errors such as IOError.
The set of errors in the ab-initio is much much broader. It includes wrong input data, segmentation
faults, problems with the resource manager, etc. The flow tries to handle the most common cases
but there's still a lot of room for improvement.
Note, in particular, that `PyFlowScheduler` will shutdown automatically in the following cases:
#. The number of python exceptions is > max_num_pyexcs
#. The number of task errors (i.e. the number of tasks whose status is S_ERROR) is > max_num_abierrs
#. The number of jobs launched becomes greater than (`safety_ratio` * total_number_of_tasks).
#. The scheduler will send an email to the user (specified by `mailto`) every `remindme_s` seconds.
If the mail cannot be sent, the scheduler will shutdown automatically.
This check prevents the scheduler from being trapped in an infinite loop.
"""
# Configuration file.
YAML_FILE = "scheduler.yml"
USER_CONFIG_DIR = os.path.join(os.path.expanduser("~"), ".abinit", "abipy")
Error = PyFlowSchedulerError
@classmethod
def autodoc(cls):
i = cls.__init__.__doc__.index("Args:")
return cls.__init__.__doc__[i + 5:]
def __init__(self, **kwargs):
"""
Args:
weeks: number of weeks to wait (DEFAULT: 0).
days: number of days to wait (DEFAULT: 0).
hours: number of hours to wait (DEFAULT: 0).
minutes: number of minutes to wait (DEFAULT: 0).
seconds: number of seconds to wait (DEFAULT: 0).
mailto: The scheduler will send an email to `mailto` every `remindme_s` seconds.
(DEFAULT: None i.e. not used).
verbose: (int) verbosity level. (DEFAULT: 0)
use_dynamic_manager: "yes" if the :class:`TaskManager` must be re-initialized from
file before launching the jobs. (DEFAULT: "no")
max_njobs_inqueue: Limit on the number of jobs that can be present in the queue. (DEFAULT: 200)
remindme_s: The scheduler will send an email to the user specified by `mailto` every `remindme_s` seconds.
(int, DEFAULT: 1 day).
max_num_pyexcs: The scheduler will exit if the number of python exceptions is > max_num_pyexcs
(int, DEFAULT: 0)
max_num_abierrs: The scheduler will exit if the number of errored tasks is > max_num_abierrs
(int, DEFAULT: 0)
safety_ratio: The scheduler will exits if the number of jobs launched becomes greater than
`safety_ratio` * total_number_of_tasks_in_flow. (int, DEFAULT: 5)
max_nlaunches: Maximum number of tasks launched in a single iteration of the scheduler.
(DEFAULT: -1 i.e. no limit)
debug: Debug level. Use 0 for production (int, DEFAULT: 0)
fix_qcritical: "yes" if the launcher should try to fix QCritical Errors (DEFAULT: "yes")
rmflow: If "yes", the scheduler will remove the flow directory if the calculation
completed successfully. (DEFAULT: "no")
killjobs_if_errors: "yes" if the scheduler should try to kill all the runnnig jobs
before exiting due to an error. (DEFAULT: "yes")
"""
# Options passed to the scheduler.
self.sched_options = AttrDict(
weeks=kwargs.pop("weeks", 0),
days=kwargs.pop("days", 0),
hours=kwargs.pop("hours", 0),
minutes=kwargs.pop("minutes", 0),
seconds=kwargs.pop("seconds", 0),
#start_date=kwargs.pop("start_date", None),
)
if all(not v for v in self.sched_options.values()):
raise self.Error("Wrong set of options passed to the scheduler.")
self.mailto = kwargs.pop("mailto", None)
self.verbose = int(kwargs.pop("verbose", 0))
self.use_dynamic_manager = as_bool(
kwargs.pop("use_dynamic_manager", False))
self.max_njobs_inqueue = kwargs.pop("max_njobs_inqueue", 200)
self.max_ncores_used = kwargs.pop("max_ncores_used", None)
self.contact_resource_manager = as_bool(
kwargs.pop("contact_resource_manager", False))
self.remindme_s = float(kwargs.pop("remindme_s", 1 * 24 * 3600))
self.max_num_pyexcs = int(kwargs.pop("max_num_pyexcs", 0))
self.max_num_abierrs = int(kwargs.pop("max_num_abierrs", 0))
self.safety_ratio = int(kwargs.pop("safety_ratio", 5))
#self.max_etime_s = kwargs.pop("max_etime_s", )
self.max_nlaunches = kwargs.pop("max_nlaunches", -1)
self.debug = kwargs.pop("debug", 0)
self.fix_qcritical = as_bool(kwargs.pop("fix_qcritical", True))
self.rmflow = as_bool(kwargs.pop("rmflow", False))
self.killjobs_if_errors = as_bool(
kwargs.pop("killjobs_if_errors", True))
self.customer_service_dir = kwargs.pop("customer_service_dir", None)
if self.customer_service_dir is not None:
self.customer_service_dir = Directory(self.customer_service_dir)
self._validate_customer_service()
if kwargs:
raise self.Error("Unknown arguments %s" % kwargs)
if not has_apscheduler:
raise RuntimeError("Install apscheduler with pip")
if has_sched_v3:
logger.warning("Using scheduler v>=3.0.0")
from apscheduler.schedulers.blocking import BlockingScheduler
self.sched = BlockingScheduler()
else:
from apscheduler.scheduler import Scheduler
self.sched = Scheduler(standalone=True)
self.nlaunch = 0
self.num_reminders = 1
# Used to keep track of the exceptions raised while the scheduler is running
self.exceptions = deque(maxlen=self.max_num_pyexcs + 10)
# Used to push additional info during the execution.
self.history = deque(maxlen=100)
@classmethod
def from_file(cls, filepath):
"""Read the configuration parameters from a Yaml file."""
with open(filepath, "rt") as fh:
return cls(**yaml.safe_load(fh))
@classmethod
def from_string(cls, s):
"""Create an istance from string s containing a YAML dictionary."""
stream = cStringIO(s)
stream.seek(0)
return cls(**yaml.safe_load(stream))
@classmethod
def from_user_config(cls):
"""
Initialize the :class:`PyFlowScheduler` from the YAML file 'scheduler.yml'.
Search first in the working directory and then in the configuration directory of abipy.
Raises:
`RuntimeError` if file is not found.
"""
# Try in the current directory.
path = os.path.join(os.getcwd(), cls.YAML_FILE)
if os.path.exists(path):
return cls.from_file(path)
# Try in the configuration directory.
path = os.path.join(cls.USER_CONFIG_DIR, cls.YAML_FILE)
if os.path.exists(path):
return cls.from_file(path)
raise cls.Error(
"Cannot locate %s neither in current directory nor in %s" %
(cls.YAML_FILE, path))
def __str__(self):
"""String representation."""
lines = [self.__class__.__name__ + ", Pid: %d" % self.pid]
app = lines.append
app("Scheduler options: %s" % str(self.sched_options))
if self.flow is not None:
app(80 * "=")
app(str(self.flow))
return "\n".join(lines)
@property
def pid(self):
"""The pid of the process associated to the scheduler."""
try:
return self._pid
except AttributeError:
self._pid = os.getpid()
return self._pid
@property
def pid_file(self):
"""
Absolute path of the file with the pid.
The file is located in the workdir of the flow
"""
return self._pid_file
@property
def flow(self):
"""`Flow`."""
try:
return self._flow
except AttributeError:
return None
@property
def num_excs(self):
"""Number of exceptions raised so far."""
return len(self.exceptions)
def get_delta_etime(self):
"""Returns a `timedelta` object representing with the elapsed time."""
return timedelta(seconds=(time.time() - self.start_time))
def add_flow(self, flow):
"""
Add an :class:`Flow` flow to the scheduler.
"""
if hasattr(self, "_flow"):
raise self.Error("Only one flow can be added to the scheduler.")
# Check if we are already using a scheduler to run this flow
flow.check_pid_file()
flow.set_spectator_mode(False)
# Build dirs and files (if not yet done)
flow.build()
with open(flow.pid_file, "wt") as fh:
fh.write(str(self.pid))
self._pid_file = flow.pid_file
self._flow = flow
def _validate_customer_service(self):
"""
Validate input parameters if customer service is on then
create directory for tarball files with correct premissions for user and group.
"""
direc = self.customer_service_dir
if not direc.exists:
mode = 0o750
print("Creating customer_service_dir %s with mode %s" %
(direc, mode))
direc.makedirs()
os.chmod(direc.path, mode)
if self.mailto is None:
raise RuntimeError(
"customer_service_dir requires mailto option in scheduler.yml")
def _do_customer_service(self):
"""
This method is called before the shutdown of the scheduler.
If customer_service is on and the flow didn't completed successfully,
a lightweight tarball file with inputs and the most important output files
is created in customer_servide_dir.
"""
if self.customer_service_dir is None: return
doit = self.exceptions or not self.flow.all_ok
doit = True
if not doit: return
prefix = os.path.basename(self.flow.workdir) + "_"
import tempfile, datetime
suffix = str(datetime.datetime.now()).replace(" ", "-")
# Remove milliseconds
i = suffix.index(".")
if i != -1: suffix = suffix[:i]
suffix += ".tar.gz"
#back = os.getcwd()
#os.chdir(self.customer_service_dir.path)
_, tmpname = tempfile.mkstemp(suffix="_" + suffix,
prefix=prefix,
dir=self.customer_service_dir.path,
text=False)
print("Dear customer,\n We are about to generate a tarball in\n %s" %
tmpname)
self.flow.make_light_tarfile(name=tmpname)
#os.chdir(back)
def start(self):
"""
Starts the scheduler in a new thread. Returns 0 if success.
In standalone mode, this method will block until there are no more scheduled jobs.
"""
self.history.append("Started on %s" % time.asctime())
self.start_time = time.time()
if not has_apscheduler:
raise RuntimeError("Install apscheduler with pip")
if has_sched_v3:
self.sched.add_job(self.callback, "interval", **self.sched_options)
else:
self.sched.add_interval_job(self.callback, **self.sched_options)
errors = self.flow.look_before_you_leap()
if errors:
self.exceptions.append(errors)
return 1
# Try to run the job immediately. If something goes wrong return without initializing the scheduler.
self._runem_all()
if self.exceptions:
self.cleanup()
self.send_email(
msg=
"Error while trying to run the flow for the first time!\n %s" %
self.exceptions)
return 1
try:
self.sched.start()
return 0
except KeyboardInterrupt:
self.shutdown(msg="KeyboardInterrupt from user")
if ask_yesno(
"Do you want to cancel all the jobs in the queue? [Y/n]"):
print("Number of jobs cancelled:", self.flow.cancel())
self.flow.pickle_dump()
return -1
def _runem_all(self):
"""
This function checks the status of all tasks,
tries to fix tasks that went unconverged, abicritical, or queuecritical
and tries to run all the tasks that can be submitted.+
"""
excs = []
flow = self.flow
# Allow to change the manager at run-time
if self.use_dynamic_manager:
from pymatgen.io.abinit.tasks import TaskManager
new_manager = TaskManager.from_user_config()
for work in flow:
work.set_manager(new_manager)
nqjobs = 0
if self.contact_resource_manager:
# This call is expensive and therefore it's optional
nqjobs = flow.get_njobs_in_queue()
if nqjobs is None:
nqjobs = 0
if flow.manager.has_queue:
logger.warning('Cannot get njobs_inqueue')
if nqjobs >= self.max_njobs_inqueue:
print("Too many jobs in the queue: %s, returning" % nqjobs)
return
if self.max_nlaunches == -1:
max_nlaunch = self.max_njobs_inqueue - nqjobs
else:
max_nlaunch = min(self.max_njobs_inqueue - nqjobs,
self.max_nlaunches)
# check status.
flow.check_status(show=False)
# This check is not perfect, we should make a list of tasks to sumbit
# and select only the subset so that we don't exceeed mac_ncores_used
# Many sections of this code should be rewritten.
#if self.max_ncores_used is not None and flow.ncores_used > self.max_ncores_used:
if self.max_ncores_used is not None and flow.ncores_allocated > self.max_ncores_used:
print("Cannot exceed max_ncores_use:d %s" % self.max_ncores_used)
return
# Try to restart the unconverged tasks
# TODO: do not fire here but prepare for fireing in rapidfire
for task in self.flow.unconverged_tasks:
try:
logger.info("Flow will try restart task %s" % task)
fired = task.restart()
if fired:
self.nlaunch += 1
max_nlaunch -= 1
if max_nlaunch == 0:
logger.info(
"Restart: too many jobs in the queue, returning")
flow.pickle_dump()
return
except task.RestartError:
excs.append(straceback())
# Temporarily disable by MG because I don't know if fix_critical works after the
# introduction of the new qadapters
# reenabled by MsS disable things that do not work at low level
# fix only prepares for restarting, and sets to ready
if self.fix_qcritical:
nfixed = flow.fix_queue_critical()
if nfixed: print("Fixed %d QCritical error(s)" % nfixed)
nfixed = flow.fix_abicritical()
if nfixed: print("Fixed %d AbiCritical error(s)" % nfixed)
# update database
flow.pickle_dump()
# Submit the tasks that are ready.
try:
nlaunch = PyLauncher(flow).rapidfire(max_nlaunch=max_nlaunch,
sleep_time=10)
self.nlaunch += nlaunch
if nlaunch:
print("[%s] Number of launches: %d" %
(time.asctime(), nlaunch))
except Exception:
excs.append(straceback())
# check status.
flow.show_status()
if excs:
logger.critical("*** Scheduler exceptions:\n *** %s" %
"\n".join(excs))
self.exceptions.extend(excs)
def callback(self):
"""The function that will be executed by the scheduler."""
try:
return self._callback()
except:
# All exceptions raised here will trigger the shutdown!
s = straceback()
self.exceptions.append(s)
# This is useful when debugging
#try:
# print("Exception in callback, will cancel all tasks")
# for task in self.flow.iflat_tasks():
# task.cancel()
#except Exception:
# pass
self.shutdown(msg="Exception raised in callback!\n" + s)
def _callback(self):
"""The actual callback."""
if self.debug:
# Show the number of open file descriptors
print(">>>>> _callback: Number of open file descriptors: %s" %
get_open_fds())
self._runem_all()
# Mission accomplished. Shutdown the scheduler.
all_ok = self.flow.all_ok
if all_ok:
return self.shutdown(
msg=
"All tasks have reached S_OK. Will shutdown the scheduler and exit"
)
# Handle failures.
err_lines = []
# Shall we send a reminder to the user?
delta_etime = self.get_delta_etime()
if delta_etime.total_seconds() > self.num_reminders * self.remindme_s:
self.num_reminders += 1
msg = (
"Just to remind you that the scheduler with pid %s, flow %s\n has been running for %s "
% (self.pid, self.flow, delta_etime))
retcode = self.send_email(msg, tag="[REMINDER]")
if retcode:
# Cannot send mail, shutdown now!
msg += (
"\nThe scheduler tried to send an e-mail to remind the user\n"
+ " but send_email returned %d. Aborting now" % retcode)
err_lines.append(msg)
#if delta_etime.total_seconds() > self.max_etime_s:
# err_lines.append("\nExceeded max_etime_s %s. Will shutdown the scheduler and exit" % self.max_etime_s)
# Too many exceptions. Shutdown the scheduler.
if self.num_excs > self.max_num_pyexcs:
msg = "Number of exceptions %s > %s. Will shutdown the scheduler and exit" % (
self.num_excs, self.max_num_pyexcs)
err_lines.append(boxed(msg))
# Paranoid check: disable the scheduler if we have submitted
# too many jobs (it might be due to some bug or other external reasons
# such as race conditions between difference callbacks!)
if self.nlaunch > self.safety_ratio * self.flow.num_tasks:
msg = "Too many jobs launched %d. Total number of tasks = %s, Will shutdown the scheduler and exit" % (
self.nlaunch, self.flow.num_tasks)
err_lines.append(boxed(msg))
# Count the number of tasks with status == S_ERROR.
if self.flow.num_errored_tasks > self.max_num_abierrs:
msg = "Number of tasks with ERROR status %s > %s. Will shutdown the scheduler and exit" % (
self.flow.num_errored_tasks, self.max_num_abierrs)
err_lines.append(boxed(msg))
# Test on the presence of deadlocks.
g = self.flow.find_deadlocks()
if g.deadlocked:
# Check the flow again so that status are updated.
self.flow.check_status()
g = self.flow.find_deadlocks()
print("deadlocked:\n", g.deadlocked, "\nrunnables:\n", g.runnables,
"\nrunning\n", g.running)
if g.deadlocked and not g.runnables and not g.running:
err_lines.append(
"No runnable job with deadlocked tasks:\n%s." %
str(g.deadlocked))
if not g.runnables and not g.running:
# Check the flow again so that status are updated.
self.flow.check_status()
g = self.flow.find_deadlocks()
if not g.runnables and not g.running:
err_lines.append(
"No task is running and cannot find other tasks to submit."
)
# Something wrong. Quit
if err_lines:
# Cancel all jobs.
if self.killjobs_if_errors:
cprint(
"killjobs_if_errors set to 'yes' in scheduler file. Will kill jobs before exiting.",
"yellow")
try:
num_cancelled = 0
for task in self.flow.iflat_tasks():
num_cancelled += task.cancel()
cprint("Killed %d tasks" % num_cancelled, "yellow")
except Exception as exc:
cprint(
"Exception while trying to kill jobs:\n%s" % str(exc),
"red")
self.shutdown("\n".join(err_lines))
return len(self.exceptions)
def cleanup(self):
"""Cleanup routine: remove the pid file and save the pickle database"""
try:
os.remove(self.pid_file)
except OSError as exc:
logger.critical("Could not remove pid_file: %s", exc)
# Save the final status of the flow.
self.flow.pickle_dump()
def shutdown(self, msg):
"""Shutdown the scheduler."""
try:
self.cleanup()
self.history.append("Completed on: %s" % time.asctime())
self.history.append("Elapsed time: %s" % self.get_delta_etime())
if self.debug:
print(">>>>> shutdown: Number of open file descriptors: %s" %
get_open_fds())
retcode = self.send_email(msg)
if self.debug:
print("send_mail retcode", retcode)
# Write file with the list of exceptions:
if self.exceptions:
dump_file = os.path.join(self.flow.workdir, "_exceptions")
with open(dump_file, "wt") as fh:
fh.writelines(self.exceptions)
fh.write("Shutdown message:\n%s" % msg)
lines = []
app = lines.append
app("Submitted on: %s" % time.ctime(self.start_time))
app("Completed on: %s" % time.asctime())
app("Elapsed time: %s" % str(self.get_delta_etime()))
if self.flow.all_ok:
app("Flow completed successfully")
else:
app("Flow %s didn't complete successfully" %
repr(self.flow.workdir))
app("use `abirun.py FLOWDIR debug` to analyze the problem.")
app("Shutdown message:\n%s" % msg)
print("")
print("\n".join(lines))
print("")
self._do_customer_service()
if self.flow.all_ok:
print("Calling flow.finalize()...")
self.flow.finalize()
#print("finalized:", self.flow.finalized)
if self.rmflow:
app("Flow directory will be removed...")
try:
self.flow.rmtree()
except Exception:
logger.warning(
"Ignoring exception while trying to remove flow dir."
)
finally:
# Shutdown the scheduler thus allowing the process to exit.
logger.debug('This should be the shutdown of the scheduler')
# Unschedule all the jobs before calling shutdown
#self.sched.print_jobs()
if not has_sched_v3:
for job in self.sched.get_jobs():
self.sched.unschedule_job(job)
#self.sched.print_jobs()
self.sched.shutdown()
# Uncomment the line below if shutdown does not work!
#os.system("kill -9 %d" % os.getpid())
def send_email(self, msg, tag=None):
"""
Send an e-mail before completing the shutdown.
Returns 0 if success.
"""
try:
return self._send_email(msg, tag)
except:
self.exceptions.append(straceback())
return -2
def _send_email(self, msg, tag):
if self.mailto is None:
return -1
header = msg.splitlines()
app = header.append
app("Submitted on: %s" % time.ctime(self.start_time))
app("Completed on: %s" % time.asctime())
app("Elapsed time: %s" % str(self.get_delta_etime()))
app("Number of errored tasks: %d" % self.flow.num_errored_tasks)
app("Number of unconverged tasks: %d" %
self.flow.num_unconverged_tasks)
strio = cStringIO()
strio.writelines("\n".join(header) + 4 * "\n")
# Add the status of the flow.
self.flow.show_status(stream=strio)
if self.exceptions:
# Report the list of exceptions.
strio.writelines(self.exceptions)
if tag is None:
tag = " [ALL OK]" if self.flow.all_ok else " [WARNING]"
return sendmail(subject=self.flow.name + tag,
text=strio.getvalue(),
mailto=self.mailto)