Celery One allows you to prevent multiple execution and queuing of celery tasks.
The project is a fork of celery-once to which I give all credits
for the initial architecture and part of this README :). The project was created to add more functionality to the original version.
celery_one
is compatible with celery_once
's previous behaviour if properly configured.
The main differences are the following:
- Option to prevent tasks with same id
- Option to get an AsyncResult if the task is already queue
- Compatibility with chords and groups
- Optional infinite timeout
- Redis connection pool
To install celery_one
with pip just run:
pip install -e git+https://github.com/szandara/celery-one.git#egg=celery_one
- Celery. Built to run with Celery 3.1. Older versions may work, but are not officially supported.
- Redis is used as a distributed locking mechanism.
To use celery_one
, your tasks need to inherit from an abstract base task called QueueOne
.
You may need to tune the following Celery configuration options...
ONE_REDIS_URL
should point towards a running Redis instance (defaults toredis://localhost:6379/0
)ONE_DEFAULT_TIMEOUT
how many seconds after a lock has been set before it should automatically timeout (defaults to 3600 seconds, or 1 hour).
from celery import Celery
from celery_one import QueueOne
from time import sleep
celery = Celery('tasks', broker='amqp://guest@localhost//')
celery.conf.ONE_REDIS_URL = 'redis://localhost:6379/0'
celery.conf.ONE_DEFAULT_TIMEOUT = 60 * 60
@celery.task(base=QueueOne)
def slow_task():
sleep(2)
return "Done!"
Behind the scenes, this overrides apply_async
and delay
. To do so, celery_one
uses Redis. Each time a task
the user calls an asynchronous task, it locks a key into Redis. If a key with a same key is pushed to the queue, its
execution is prevented by either raising an AlreadyQueued
exception or returning the AsyncResult of the already running task.
The key of the task is generated from its name and parameters.
example.delay(10)
example.delay(10)
Traceback (most recent call last):
..
AlreadyQueued()
result1 = example.apply_async(args=(10,))
result2 = example.apply_async(args=(10,), one_options={'fail':False}) # No failure here
assert result1.id == result2.id
result = example.apply_async(args=(10,))
Traceback (most recent call last):
..
AlreadyQueued()
If this option is set, the key will be generated using the id of the task. This can be useful when working with meaningful task ids or when the arguments are not necessarily indication of different tasks.
@celery.task(base=QueueOne, one_options={'use_id':True})
def slow_task_no_fail(a):
print("Running")
sleep(10)
return "Done: " + str(a)
result1 = slow_task_no_fail.apply_async(args=(10), task_id=id1)
result2 = slow_task_no_fail.apply_async(args=(12), task_id=id1)
Traceback (most recent call last):
..
AlreadyQueued()
Optionally, instead of raising an AlreadyQueued
exception, the task can return an AsyncResult.
To do so, set the option in the celery task or directly in the apply_async
call.
@celery.task(base=QueueOne, one_options={'fail':False})
def slow_task_no_fail():
print("Running")
sleep(2)
return "Done!"
result1 = slow_task_no_fail.apply_async(args=(10))
result2 = slow_task_no_fail.apply_async(args=(10))
print(result1.get())
print(result2.get())
Output:
Running
Done!
Done!
By default QueueOne
creates a lock based on the task's name and its arguments and values.
Take for example, the following task below...
@celery.task(base=QueueOne)
def slow_add(a, b):
sleep(2)
return a + b
Running the task with different arguments will default to checking against different locks.
slow_add(1, 1)
slow_add(1, 2)
If you want to specify locking based on a subset, or no arguments you can adjust the keys celery_one
looks at in the task's options with one_options={'keys': [..]}
@celery.task(base=QueueOne, one_options={'keys': ['a']})
def slow_add(a, b):
sleep(30)
return a + b
example.delay(1, 1)
# Checks if any tasks are running with the `a=1`
example.delay(1, 2)
Traceback (most recent call last):
..
AlreadyQueued()
example.delay(2, 2)
@celery.task(base=QueueOne, one_options={'keys': []})
def slow_add(a, b):
sleep(30)
return a + b
# Will enforce only one task can run, no matter what arguments.
example.delay(1, 1)
example.delay(2, 2)
Traceback (most recent call last):
..
AlreadyQueued()
As a fall back, celery_one
will clear a lock after 60 minutes.
This is set globally in Celery's configuration with ONE_DEFAULT_TIMEOUT
but can be set for individual tasks using...
@celery.task(base=QueueOne, one_options={'timeout': 60 * 60 * 10})
def long_running_task():
sleep(60 * 60 * 3)
timeout
can also be set to None, causing celery_one
to prevent adding same tasks until the already running one
is complete. NOTE: This might result in dangerous behaviors such as deadlocks or failing task executions. Use with care!
By default, the lock is removed after the task has executed (using celery's after_return). This behaviour can be changed setting the task's option unlock_before_run
. When set to True
, the lock will be removed just before executing the task.
Caveat: any retry of the task won't re-enable the lock!
@celery.task(base=QueueOne, one_options={'unlock_before_run': True})
def slow_task():
sleep(30)
return "Done!"
- Tests are run against Python 2.7 and 3.3. Other versions may work, but are not officially supported.
Contributions are welcome, and they are greatly appreciated! See contributing guide for more details.