This document describes an older version of Celery (2.1). For the latest stable version please go here.

Multiprocessing Worker - celery.worker

The Multiprocessing Worker Server

class celery.worker.WorkController(concurrency=None, logfile=None, loglevel=None, send_events=False, hostname=None, ready_callback=<function noop at 0x4741320>, embed_clockservice=False, pool_cls='celery.concurrency.processes.TaskPool', listener_cls='celery.worker.listener.CarrotListener', mediator_cls='celery.worker.controllers.Mediator', eta_scheduler_cls='celery.utils.timer2.Timer', schedule_filename='celerybeat-schedule', task_time_limit=None, task_soft_time_limit=None, max_tasks_per_child=None, pool_putlocks=True, disable_rate_limits=False, db=None, scheduler_cls='celery.beat.PersistentScheduler')

Executes tasks waiting in the task queue.

Parameters:

• concurrency – see concurrency.
• logfile – see logfile.
• loglevel – see loglevel.
• embed_clockservice – see embed_clockservice.
• send_events – see send_events.

concurrency

The number of simultaneous processes doing work (default: conf.CELERYD_CONCURRENCY)

loglevel

The loglevel used (default: logging.INFO)

logfile

The logfile used, if no logfile is specified it uses stderr (default: celery.conf.CELERYD_LOG_FILE).

embed_clockservice

If True, celerybeat is embedded, running in the main worker process as a thread.

send_events

Enable the sending of monitoring events, these events can be captured by monitors (celerymon).

logger

The logging.Logger instance used for logging.

pool

The multiprocessing.Pool instance used.

ready_queue

The Queue.Queue that holds tasks ready for immediate processing.

schedule_controller

Instance of celery.worker.controllers.ScheduleController.

mediator

Instance of celery.worker.controllers.Mediator.

listener

Instance of CarrotListener.

concurrency = 0

logfile = None

loglevel = 40

on_timer_error(exc_info)

on_timer_tick(delay)

process_task(wrapper)

Process task by sending it to the pool of workers.

start()

Starts the workers main loop.

stop()

Graceful shutdown of the worker server.

terminate()

Not so graceful shutdown of the worker server.

celery.worker.process_initializer(hostname)

Initializes the process so it can be used to process tasks.

Used for multiprocessing environments.