well, after all the time you spent after it, it's quite a bold statement.
Current scheduler (in trunk there has been a problem, attached the correct
one) reassign "RUNNING" tasks if a worker is killed in action, but you may
want to check if in scheduler_worker table you have entries for dead
workers too (there should be only active schedulers on it).
PS: bugs get eventually solved, if you can point to some of them :D
PS 2 : I use scheduler in production both for "one-time-only" tasks and for
recurring ones, and never incurred in such problems.
--
#!/usr/bin/env python
# -*- coding: utf-8 -*-
USAGE = """
## Example
For any existing app
Create File: app/models/scheduler.py ======
from gluon.scheduler import Scheduler
def demo1(*args,**vars):
print 'you passed args=%s and vars=%s' % (args, vars)
return 'done!'
def demo2():
1/0
scheduler = Scheduler(db,dict(demo1=demo1,demo2=demo2))
## run worker nodes with:
cd web2py
python web2py.py -K myapp
or
python gluon/scheduler.py -u sqlite://storage.sqlite \
-f applications/myapp/databases/ \
-t mytasks.py
(-h for info)
python scheduler.py -h
## schedule jobs using
http://127.0.0.1:8000/scheduler/appadmin/insert/db/scheduler_task
## monitor scheduled jobs
http://127.0.0.1:8000/scheduler/appadmin/select/db?query=db.scheduler_task.id>0
## view completed jobs
http://127.0.0.1:8000/scheduler/appadmin/select/db?query=db.scheduler_run.id>0
## view workers
http://127.0.0.1:8000/scheduler/appadmin/select/db?query=db.scheduler_worker.id>0
## To install the scheduler as a permanent daemon on Linux (w/ Upstart), put the
## following into /etc/init/web2py-scheduler.conf:
## (This assumes your web2py instance is installed in <user>'s home directory,
## running as <user>, with app <myapp>, on network interface eth0.)
description "web2py task scheduler"
start on (local-filesystems and net-device-up IFACE=eth0)
stop on shutdown
respawn limit 8 60 # Give up if restart occurs 8 times in 60 seconds.
exec sudo -u <user> python /home/<user>/web2py/web2py.py -K <myapp>
respawn
## You can then start/stop/restart/check status of the daemon with:
sudo start web2py-scheduler
sudo stop web2py-scheduler
sudo restart web2py-scheduler
sudo status web2py-scheduler
"""
import os
import time
import multiprocessing
import sys
import cStringIO
import threading
import traceback
import signal
import socket
import datetime
import logging
import optparse
if 'WEB2PY_PATH' in os.environ:
sys.path.append(os.environ['WEB2PY_PATH'])
else:
os.environ['WEB2PY_PATH'] = os.getcwd()
if not os.environ['WEB2PY_PATH'] in sys.path:
sys.path.append(os.environ['WEB2PY_PATH'])
try:
from gluon.contrib.simplejson import loads, dumps
except:
from simplejson import loads, dumps
from gluon import DAL, Field, IS_NOT_EMPTY, IS_IN_SET
from gluon.utils import web2py_uuid
QUEUED = 'QUEUED'
ASSIGNED = 'ASSIGNED'
RUNNING = 'RUNNING'
COMPLETED = 'COMPLETED'
FAILED = 'FAILED'
TIMEOUT = 'TIMEOUT'
STOPPED = 'STOPPED'
ACTIVE = 'ACTIVE'
INACTIVE = 'INACTIVE'
DISABLED = 'DISABLED'
EXPIRED = 'EXPIRED'
SECONDS = 1
HEARTBEAT = 3*SECONDS
MAXHIBERNATION = 10
class Task(object):
def __init__(self,app,function,timeout,args='[]',vars='{}',**kwargs):
logging.debug(' new task allocated: %s.%s' % (app,function))
self.app = app
self.function = function
self.timeout = timeout
self.args = args # json
self.vars = vars # json
self.__dict__.update(kwargs)
def __str__(self):
return '<Task: %s>' % self.function
class TaskReport(object):
def __init__(self,status,result=None,output=None,tb=None):
logging.debug(' new task report: %s' % status)
if tb:
logging.debug(' traceback: %s' % tb)
else:
logging.debug(' result: %s' % result)
self.status = status
self.result = result
self.output = output
self.tb = tb
def __str__(self):
return '<TaskReport: %s>' % self.status
def demo_function(*argv,**kwargs):
""" test function """
for i in range(argv[0]):
print 'click',i
time.sleep(1)
return 'done'
#the two functions below deal with simplejson decoding as unicode, esp for the dict decode
#and subsequent usage as function Keyword arguments unicode variable names won't work!
#borrowed from http://stackoverflow.com/questions/956867/how-to-get-string-objects-instead-unicode-ones-from-json-in-python
def _decode_list(lst):
newlist = []
for i in lst:
if isinstance(i, unicode):
i = i.encode('utf-8')
elif isinstance(i, list):
i = _decode_list(i)
newlist.append(i)
return newlist
def _decode_dict(dct):
newdict = {}
for k, v in dct.iteritems():
if isinstance(k, unicode):
k = k.encode('utf-8')
if isinstance(v, unicode):
v = v.encode('utf-8')
elif isinstance(v, list):
v = _decode_list(v)
newdict[k] = v
return newdict
def executor(queue,task):
""" the background process """
logging.debug(' task started')
stdout, sys.stdout = sys.stdout, cStringIO.StringIO()
try:
if task.app:
os.chdir(os.environ['WEB2PY_PATH'])
from gluon.shell import env
from gluon.dal import BaseAdapter
from gluon import current
level = logging.getLogger().getEffectiveLevel()
logging.getLogger().setLevel(logging.WARN)
_env = env(task.app,import_models=True)
logging.getLogger().setLevel(level)
scheduler = current._scheduler
scheduler_tasks = current._scheduler.tasks
_function = scheduler_tasks[task.function]
globals().update(_env)
args = loads(task.args)
vars = loads(task.vars, object_hook=_decode_dict)
result = dumps(_function(*args,**vars))
else:
### for testing purpose only
result = eval(task.function)(
*loads(task.args, object_hook=_decode_dict),
**loads(task.vars, object_hook=_decode_dict))
stdout, sys.stdout = sys.stdout, stdout
queue.put(TaskReport(COMPLETED, result,stdout.getvalue()))
except BaseException,e:
sys.stdout = stdout
tb = traceback.format_exc()
queue.put(TaskReport(FAILED,tb=tb))
class MetaScheduler(threading.Thread):
def __init__(self):
threading.Thread.__init__(self)
self.process = None # the backround process
self.have_heartbeat = True # set to False to kill
self.empty_runs = 0
def async(self,task):
"""
starts the background process and returns:
('ok',result,output)
('error',exception,None)
('timeout',None,None)
('terminated',None,None)
"""
queue = multiprocessing.Queue(maxsize=1)
p = multiprocessing.Process(target=executor,args=(queue,task))
self.process = p
logging.debug(' task starting')
p.start()
try:
p.join(task.timeout)
except:
p.terminate()
p.join()
self.have_heartbeat = False
logging.debug(' task stopped')
return TaskReport(STOPPED)
if p.is_alive():
p.terminate()
p.join()
logging.debug(' task timeout')
return TaskReport(TIMEOUT)
elif queue.empty():
self.have_heartbeat = False
logging.debug(' task stopped')
return TaskReport(STOPPED)
else:
logging.debug(' task completed or failed')
return queue.get()
def die(self):
logging.info('die!')
self.have_heartbeat = False
self.terminate_process()
def terminate_process(self):
try:
self.process.terminate()
except:
pass # no process to terminate
def run(self):
""" the thread that sends heartbeat """
counter = 0
while self.have_heartbeat:
self.send_heartbeat(counter)
counter += 1
def start_heartbeats(self):
self.start()
def send_heartbeat(self,counter):
print 'thum'
time.sleep(1)
def pop_task(self):
return Task(
app = None,
function = 'demo_function',
timeout = 7,
args = '[2]',
vars = '{}')
def report_task(self,task,task_report):
print 'reporting task'
pass
def sleep(self):
pass
def loop(self):
try:
self.start_heartbeats()
while True and self.have_heartbeat:
logging.debug('looping...')
task = self.pop_task()
if task:
self.empty_runs = 0
self.report_task(task,self.async(task))
else:
self.empty_runs += 1
logging.debug('sleeping...')
if self.max_empty_runs <> 0:
logging.debug('empty runs %s/%s', self.empty_runs, self.max_empty_runs)
if self.empty_runs >= self.max_empty_runs:
logging.info('empty runs limit reached, killing myself')
self.die()
self.sleep()
except KeyboardInterrupt:
self.die()
TASK_STATUS = (QUEUED, RUNNING, COMPLETED, FAILED, TIMEOUT, STOPPED, EXPIRED)
RUN_STATUS = (RUNNING, COMPLETED, FAILED, TIMEOUT, STOPPED)
WORKER_STATUS = (ACTIVE,INACTIVE,DISABLED)
class TYPE(object):
"""
validator that check whether field is valid json and validate its type
"""
def __init__(self,myclass=list,parse=False):
self.myclass = myclass
self.parse=parse
def __call__(self,value):
from gluon import current
try:
obj = loads(value)
except:
return (value,current.T('invalid json'))
else:
if isinstance(obj,self.myclass):
if self.parse:
return (obj,None)
else:
return (value,None)
else:
return (value,current.T('Not of type: %s') % self.myclass)
class Scheduler(MetaScheduler):
def __init__(self,db,tasks={},migrate=True,
worker_name=None,group_names=None,heartbeat=HEARTBEAT,max_empty_runs=0):
MetaScheduler.__init__(self)
self.db = db
self.db_thread = None
self.tasks = tasks
self.group_names = group_names or ['main']
self.heartbeat = heartbeat
self.worker_name = worker_name or socket.gethostname()+'#'+str(web2py_uuid())
self.worker_status = RUNNING, 1 #tuple containing status as recorded in
#the table, plus a boost parameter for
#hibernation (i.e. when someone stop the
#worker acting on the scheduler_worker table)
self.max_empty_runs = max_empty_runs
self.is_a_ticker = False
self.do_assign_tasks = False
from gluon import current
current._scheduler = self
self.define_tables(db,migrate=migrate)
def define_tables(self,db,migrate):
from gluon import current
logging.debug('defining tables (migrate=%s)' % migrate)
now = datetime.datetime.now()
db.define_table(
'scheduler_task',
Field('application_name',requires=IS_NOT_EMPTY(),
default=None,writable=False),
Field('task_name',requires=IS_NOT_EMPTY()),
Field('group_name',default='main',writable=False),
Field('status',requires=IS_IN_SET(TASK_STATUS),
default=QUEUED,writable=False),
Field('function_name',
requires=IS_IN_SET(sorted(self.tasks.keys()))),
Field('args','text',default='[]',requires=TYPE(list)),
Field('vars','text',default='{}',requires=TYPE(dict)),
Field('enabled','boolean',default=True),
Field('start_time','datetime',default=now),
Field('next_run_time','datetime',default=now),
Field('stop_time','datetime'),
Field('repeats','integer',default=1,comment="0=unlimted"),
Field('period','integer',default=60,comment='seconds'),
Field('timeout','integer',default=60,comment='seconds'),
Field('times_run','integer',default=0,writable=False),
Field('last_run_time','datetime',writable=False,readable=False),
Field('assigned_worker_name',default='',writable=False),
migrate=migrate,format='%(task_name)s')
if hasattr(current,'request'):
db.scheduler_task.application_name.default=current.request.application
db.define_table(
'scheduler_run',
Field('scheduler_task','reference scheduler_task'),
Field('status',requires=IS_IN_SET(RUN_STATUS)),
Field('start_time','datetime'),
Field('stop_time','datetime'),
Field('output','text'),
Field('result','text'),
Field('traceback','text'),
Field('worker_name',default=self.worker_name),
migrate=migrate)
db.define_table(
'scheduler_worker',
Field('worker_name'),
Field('first_heartbeat','datetime'),
Field('last_heartbeat','datetime'),
Field('status',requires=IS_IN_SET(WORKER_STATUS)),
Field('is_ticker', 'boolean', default=False, writable=False),
Field('group_names', 'list:string', default=self.group_names),
migrate=migrate)
db.commit()
def loop(self,worker_name=None):
try:
self.start_heartbeats()
while True and self.have_heartbeat:
if self.worker_status[0] == DISABLED:
logging.debug('Someone stopped me, sleeping until better times come (%s)' % self.worker_status[1])
self.sleep()
continue
logging.debug('looping...')
task = self.pop_task()
if task:
self.empty_runs = 0
self.report_task(task,self.async(task))
else:
self.empty_runs += 1
logging.debug('sleeping...')
if self.max_empty_runs <> 0:
logging.debug('empty runs %s/%s', self.empty_runs, self.max_empty_runs)
if self.empty_runs >= self.max_empty_runs:
logging.info('empty runs limit reached, killing myself')
self.die()
self.sleep()
except KeyboardInterrupt:
self.die()
def pop_task(self):
now = datetime.datetime.now()
db, ts = self.db, self.db.scheduler_task
if self.is_a_ticker and self.do_assign_tasks:
#I'm a ticker, and 5 loops passed without reassigning tasks, let's do
#that and loop again
db.commit() #?don't know if it's useful, let's be completely sure
while True:
try:
self.assign_tasks()
db.commit()
break
except:
db.rollback()
logging.error('TICKER: error assigning tasks')
return None
grabbed = db(ts.assigned_worker_name==self.worker_name)\
(ts.status==ASSIGNED)
task = grabbed.select(limitby=(0,1), orderby=ts.next_run_time).first()
if task:
task.update_record(status=RUNNING,last_run_time=now)
#noone will touch my task!
db.commit()
logging.debug(' work to do %s' % task.id)
else:
logging.debug('nothing to do')
return None
next_run_time = task.last_run_time + datetime.timedelta(seconds=task.period)
times_run = task.times_run + 1
if times_run < task.repeats or task.repeats==0:
run_again = True
else:
run_again = False
logging.debug(' new scheduler_run record')
while True:
try:
run_id = db.scheduler_run.insert(
scheduler_task = task.id,
status=RUNNING,
start_time=now,
worker_name=self.worker_name)
db.commit()
break
except:
db.rollback()
logging.info('new task %(id)s "%(task_name)s" %(application_name)s.%(function_name)s' % task)
return Task(
app = task.application_name,
function = task.function_name,
timeout = task.timeout,
args = task.args, #in json
vars = task.vars, #in json
task_id = task.id,
run_id = run_id,
run_again = run_again,
next_run_time=next_run_time,
times_run = times_run,
stop_time = task.stop_time)
def report_task(self,task,task_report):
logging.debug(' recording task report in db (%s)' % task_report.status)
db = self.db
db(db.scheduler_run.id==task.run_id).update(
status = task_report.status,
stop_time = datetime.datetime.now(),
result = task_report.result,
output = task_report.output,
traceback = task_report.tb)
is_expired = task.stop_time and task.next_run_time > task.stop_time and True or False
status = (task.run_again and is_expired and EXPIRED
or task.run_again and not is_expired and QUEUED or COMPLETED)
if task_report.status == COMPLETED:
d = dict(status = status,
next_run_time = task.next_run_time,
times_run = task.times_run)
#I'd like to know who worked my task, reviewing some logs...
#,assigned_worker_name = '')
else:
d = dict(
#same as before...
#assigned_worker_name = '',
status = {'FAILED':'FAILED',
'TIMEOUT':'TIMEOUT',
'STOPPED':'QUEUED'}[task_report.status])
db(db.scheduler_task.id==task.task_id)\
(db.scheduler_task.status==RUNNING).update(**d)
db.commit()
logging.info('task completed (%s)' % task_report.status)
def adj_hibernation(self):
if self.worker_status[0] == DISABLED:
hibernation = self.worker_status[1] + 1 if self.worker_status[1] < MAXHIBERNATION else MAXHIBERNATION
self.worker_status = DISABLED, hibernation
def send_heartbeat(self,counter):
if not self.db_thread:
logging.debug('thread building own DAL object')
self.db_thread = DAL(self.db._uri,folder = self.db._adapter.folder)
self.define_tables(self.db_thread,migrate=False)
try:
db = self.db_thread
sw, st = db.scheduler_worker, db.scheduler_task
now = datetime.datetime.now()
expiration = now-datetime.timedelta(seconds=self.heartbeat*3)
departure = now-datetime.timedelta(seconds=self.heartbeat*3*MAXHIBERNATION)
# record heartbeat
mybackedstatus = db(sw.worker_name==self.worker_name).select().first()
if not mybackedstatus:
sw.insert(status = ACTIVE,worker_name = self.worker_name,
first_heartbeat = now,last_heartbeat = now,
group_names = self.group_names)
self.worker_status = ACTIVE, 1 #activating the process
else:
if mybackedstatus.status == DISABLED:
self.worker_status = DISABLED, self.worker_status[1]#keep sleeping
if self.worker_status[1] == MAXHIBERNATION:
logging.debug('........recording heartbeat')
db(sw.worker_name==self.worker_name).update(
last_heartbeat = now)
else:
logging.debug('........recording heartbeat')
db(sw.worker_name==self.worker_name).update(
last_heartbeat = now, status = ACTIVE)
self.worker_status = ACTIVE, 1 #re-activating the process
self.do_assign_tasks = False
if counter % 5 == 0:
try:
# delete inactive workers
logging.debug(' freeing workers that have not sent heartbeat')
inactive_workers = db(
((sw.last_heartbeat<expiration) & (sw.status == ACTIVE)) |
((sw.last_heartbeat<departure) & (sw.status == DISABLED))
)
db(st.assigned_worker_name.belongs(
inactive_workers._select(sw.worker_name)))\
(st.status == RUNNING)\
.update(assigned_worker_name='',status=QUEUED)
inactive_workers.delete()
self.is_a_ticker = self.being_a_ticker()
if self.worker_status[0] <> DISABLED:
self.do_assign_tasks = True
except:
pass
db.commit()
except:
db.rollback()
self.adj_hibernation()
self.sleep()
def being_a_ticker(self):
db = self.db_thread
sw = db.scheduler_worker
ticker = db((sw.worker_name <> self.worker_name) & (sw.is_ticker == True) & (sw.status == ACTIVE)).select().first()
if not ticker:
db(sw.worker_name == self.worker_name).update(is_ticker = True)
db(sw.worker_name <> self.worker_name).update(is_ticker = False)
logging.info("TICKER: I'm a ticker (%s)" % self.worker_name)
return True
else:
logging.info("%s is a ticker, I'm a poor worker" % ticker.worker_name)
return False
def assign_tasks(self):
db = self.db
sw, ts = db.scheduler_worker, db.scheduler_task
now = datetime.datetime.now()
all_workers = db(sw.status <> DISABLED).select()
#build workers as dict of groups
wkgroups = {}
for w in all_workers:
group_names = w.group_names
for gname in group_names:
if gname not in wkgroups:
wkgroups[gname] = dict(workers=[{'name' : w.worker_name, 'c' : 0}])
else:
wkgroups[gname]['workers'].append({'name' : w.worker_name, 'c' : 0})
#set queued tasks that expired between "runs" (i.e., you turned off)
#the scheduler and then it wasn't expired, but now it is
db(ts.status.belongs((QUEUED,ASSIGNED)))(ts.stop_time<now).update(status=EXPIRED)
all_available = db(ts.status.belongs((QUEUED,ASSIGNED)))\
((ts.times_run<ts.repeats)|(ts.repeats==0))\
(ts.start_time<=now)\
((ts.stop_time==None) | (ts.stop_time>now))\
(ts.next_run_time<=now)\
(ts.enabled==True)
limit = len(all_workers) * 50
#if there are a moltitude of tasks, let's assign a maximum of 50 tasks per worker.
#this can be adjusted with some added intelligence (like esteeming how many tasks will
#a worker complete before the ticker reassign them around, but the gain is quite small
#50 is quite a sweet spot also for fast tasks, with sane heartbeat values
#NB: ticker reassign tasks every 5 cycles, so if a worker completes his 50 tasks in less
#than heartbeat*5 seconds, it won't pick new tasks until heartbeat*5 seconds pass.
tasks = all_available.select(limitby=(0,limit), orderby=ts.next_run_time)
#everything until now is going fine. If a worker is currently elaborating a long task,
#all other tasks assigned to him needs to be reassigned "freely" to other workers, that may be free.
#this shuffles up things a bit, in order to maintain the idea of a semi-linear scalability
#let's freeze it up
db.commit()
#let's break up the queue evenly among workers
for task in tasks:
gname = task.group_name
ws = wkgroups.get(gname)
if ws:
counter = 0
myw = 0
for i, w in enumerate(ws['workers']):
if w['c'] < counter:
myw = i
counter = w['c']
task.update_record(status=ASSIGNED,
assigned_worker_name=wkgroups[gname]['workers'][myw]['name'])
wkgroups[gname]['workers'][myw]['c'] += 1
db.commit()
#I didn't report tasks but I'm working nonetheless!!!!
if len(tasks) > 0:
self.empty_runs = 0
logging.info('TICKER: workers are %s' % len(all_workers))
logging.info('TICKER: tasks are %s' % len(tasks))
def sleep(self):
time.sleep(self.heartbeat*self.worker_status[1]) # should only sleep until next available task
def main():
"""
allows to run worker without python web2py.py .... by simply python this.py
"""
parser = optparse.OptionParser()
parser.add_option(
"-w", "--worker_name", dest="worker_name", default=None,
help="start a worker with name")
parser.add_option(
"-b", "--heartbeat",dest="heartbeat", default = 10,
type='int', help="heartbeat time in seconds (default 10)")
parser.add_option(
"-L", "--logger_level",dest="logger_level",
default=30,
type='int',
help="set debug output level (0-100, 0 means all, 100 means none;default is 30)")
parser.add_option("-E", "--empty-runs",
dest="max_empty_runs",
type='int',
default = 0,
help="max loops with no grabbed tasks permitted (0 for never check)")
parser.add_option(
"-g", "--group_names",dest="group_names",
default = 'main',
help="comma separated list of groups to be picked by the worker")
parser.add_option(
"-f", "--db_folder",dest="db_folder",
default = '/Users/mdipierro/web2py/applications/scheduler/databases',
help="location of the dal database folder")
parser.add_option(
"-u", "--db_uri",dest="db_uri",
default = 'sqlite://storage.sqlite',
help="database URI string (web2py DAL syntax)")
parser.add_option(
"-t", "--tasks",dest="tasks",default=None,
help="file containing task files, must define" + \
"tasks = {'task_name':(lambda: 'output')} or similar set of tasks")
(options, args) = parser.parse_args()
if not options.tasks or not options.db_uri:
print USAGE
if options.tasks:
path,filename = os.path.split(options.tasks)
if filename.endswith('.py'):
filename = filename[:-3]
sys.path.append(path)
print 'importing tasks...'
tasks = __import__(filename, globals(), locals(), [], -1).tasks
print 'tasks found: '+', '.join(tasks.keys())
else:
tasks = {}
group_names = [x.strip() for x in options.group_names.split(',')]
logging.getLogger().setLevel(options.logger_level)
print 'groups for this worker: '+', '.join(group_names)
print 'connecting to database in folder: ' + options.db_folder or './'
print 'using URI: '+options.db_uri
db = DAL(options.db_uri,folder=options.db_folder)
print 'instantiating scheduler...'
scheduler=Scheduler(db = db,
worker_name = options.worker_name,
tasks = tasks,
migrate = True,
group_names = group_names,
heartbeat = options.heartbeat,
max_empty_runs = options.max_empty_runs)
print 'starting main worker loop...'
scheduler.loop()
if __name__=='__main__':
main()
