faq.rst 28 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915
  1. .. _faq:
  2. ============================
  3. Frequently Asked Questions
  4. ============================
  5. .. contents::
  6. :local:
  7. .. _faq-general:
  8. General
  9. =======
  10. .. _faq-when-to-use:
  11. What kinds of things should I use Celery for?
  12. ---------------------------------------------
  13. **Answer:** `Queue everything and delight everyone`_ is a good article
  14. describing why you would use a queue in a web context.
  15. .. _`Queue everything and delight everyone`:
  16. http://decafbad.com/blog/2008/07/04/queue-everything-and-delight-everyone
  17. These are some common use cases:
  18. * Running something in the background. For example, to finish the web request
  19. as soon as possible, then update the users page incrementally.
  20. This gives the user the impression of good performance and "snappiness", even
  21. though the real work might actually take some time.
  22. * Running something after the web request has finished.
  23. * Making sure something is done, by executing it asynchronously and using
  24. retries.
  25. * Scheduling periodic work.
  26. And to some degree:
  27. * Distributed computing.
  28. * Parallel execution.
  29. .. _faq-misconceptions:
  30. Misconceptions
  31. ==============
  32. .. _faq-loc:
  33. Does Celery really consist of 50.000 lines of code?
  34. ---------------------------------------------------
  35. **Answer:** No, this and similarly large numbers have
  36. been reported at various locations.
  37. The numbers as of this writing are:
  38. - core: 7,141 lines of code.
  39. - tests: 14,209 lines.
  40. - backends, contrib, compat utilities: 9,032 lines.
  41. Lines of code is not a useful metric, so
  42. even if Celery did consist of 50k lines of code you would not
  43. be able to draw any conclusions from such a number.
  44. Does Celery have many dependencies?
  45. -----------------------------------
  46. A common criticism is that Celery uses too many dependencies.
  47. The rationale behind such a fear is hard to imagine, especially considering
  48. code reuse as the established way to combat complexity in modern software
  49. development, and that the cost of adding dependencies is very low now
  50. that package managers like pip and PyPI makes the hassle of installing
  51. and maintaining dependencies a thing of the past.
  52. Celery has replaced several dependencies along the way, and
  53. the current list of dependencies are:
  54. celery
  55. ~~~~~~
  56. - `kombu`_
  57. Kombu is part of the Celery ecosystem and is the library used
  58. to send and receive messages. It is also the library that enables
  59. us to support many different message brokers. It is also used by the
  60. OpenStack project, and many others, validating the choice to separate
  61. it from the Celery codebase.
  62. .. _`kombu`: http://pypi.python.org/pypi/kombu
  63. - `billiard`_
  64. Billiard is a fork of the Python multiprocessing module containing
  65. many performance and stability improvements. It is an eventual goal
  66. that these improvements will be merged back into Python one day.
  67. It is also used for compatibility with older Python versions
  68. that doesn't come with the multiprocessing module.
  69. .. _`billiard`: http://pypi.python.org/pypi/billiard
  70. - `pytz`
  71. The pytz module provides timezone definitions and related tools.
  72. .. _`pytz`: http://pypi.python.org/pypi/pytz
  73. django-celery
  74. ~~~~~~~~~~~~~
  75. If you use django-celery then you don't have to install celery separately,
  76. as it will make sure that the required version is installed.
  77. django-celery does not have any other dependencies.
  78. kombu
  79. ~~~~~
  80. Kombu depends on the following packages:
  81. - `amqplib`_
  82. The underlying pure-Python amqp client implementation. AMQP being the default
  83. broker it is a natural dependency.
  84. .. _`amqplib`: http://pypi.python.org/pypi/amqplib
  85. - `anyjson`_
  86. anyjson is an utility library to select the best possible
  87. JSON implementation.
  88. .. _`anyjson`: http://pypi.python.org/pypi/anyjson
  89. .. note::
  90. For compatibility reasons additional packages may be installed
  91. if you are running on older Python versions,
  92. for example Python 2.6 depends on the ``importlib``,
  93. and ``ordereddict`` libraries.
  94. Also, to handle the dependencies for popular configuration
  95. choices Celery defines a number of "bundle" packages,
  96. see :ref:`bundles`.
  97. .. _faq-heavyweight:
  98. Is Celery heavy-weight?
  99. -----------------------
  100. Celery poses very little overhead both in memory footprint and
  101. performance.
  102. But please note that the default configuration is not optimized for time nor
  103. space, see the :ref:`guide-optimizing` guide for more information.
  104. .. _faq-serializion-is-a-choice:
  105. Is Celery dependent on pickle?
  106. ------------------------------
  107. **Answer:** No.
  108. Celery can support any serialization scheme and has built-in support for
  109. JSON, YAML, Pickle and msgpack. Also, as every task is associated with a
  110. content type, you can even send one task using pickle, and another using JSON.
  111. The default serialization format is pickle simply because it is
  112. convenient (it supports sending complex Python objects as task arguments).
  113. If you need to communicate with other languages you should change
  114. to a serialization format that is suitable for that.
  115. You can set a global default serializer, the default serializer for a
  116. particular Task, or even what serializer to use when sending a single task
  117. instance.
  118. .. _faq-is-celery-for-django-only:
  119. Is Celery for Django only?
  120. --------------------------
  121. **Answer:** No.
  122. Celery does not depend on Django anymore. To use Celery with Django you have
  123. to use the `django-celery`_ package.
  124. .. _`django-celery`: http://pypi.python.org/pypi/django-celery
  125. .. _faq-is-celery-for-rabbitmq-only:
  126. Do I have to use AMQP/RabbitMQ?
  127. -------------------------------
  128. **Answer**: No.
  129. You can also use Redis, Beanstalk, CouchDB, MongoDB or an SQL database,
  130. see :ref:`brokers`.
  131. These "virtual transports" may have limited broadcast and event functionality.
  132. For example remote control commands only works with AMQP and Redis.
  133. Redis or a database won't perform as well as
  134. an AMQP broker. If you have strict reliability requirements you are
  135. encouraged to use RabbitMQ or another AMQP broker. Redis/database also use
  136. polling, so they are likely to consume more resources. However, if you for
  137. some reason are not able to use AMQP, feel free to use these alternatives.
  138. They will probably work fine for most use cases, and note that the above
  139. points are not specific to Celery; If using Redis/database as a queue worked
  140. fine for you before, it probably will now. You can always upgrade later
  141. if you need to.
  142. .. _faq-is-celery-multilingual:
  143. Is Celery multilingual?
  144. ------------------------
  145. **Answer:** Yes.
  146. :mod:`~celery.bin.celeryd` is an implementation of Celery in Python. If the
  147. language has an AMQP client, there shouldn't be much work to create a worker
  148. in your language. A Celery worker is just a program connecting to the broker
  149. to process messages.
  150. Also, there's another way to be language independent, and that is to use REST
  151. tasks, instead of your tasks being functions, they're URLs. With this
  152. information you can even create simple web servers that enable preloading of
  153. code. See: `User Guide: Remote Tasks`_.
  154. .. _`User Guide: Remote Tasks`:
  155. http://celery.github.com/celery/userguide/remote-tasks.html
  156. .. _faq-troubleshooting:
  157. Troubleshooting
  158. ===============
  159. .. _faq-mysql-deadlocks:
  160. MySQL is throwing deadlock errors, what can I do?
  161. -------------------------------------------------
  162. **Answer:** MySQL has default isolation level set to `REPEATABLE-READ`,
  163. if you don't really need that, set it to `READ-COMMITTED`.
  164. You can do that by adding the following to your :file:`my.cnf`::
  165. [mysqld]
  166. transaction-isolation = READ-COMMITTED
  167. For more information about InnoDB`s transaction model see `MySQL - The InnoDB
  168. Transaction Model and Locking`_ in the MySQL user manual.
  169. (Thanks to Honza Kral and Anton Tsigularov for this solution)
  170. .. _`MySQL - The InnoDB Transaction Model and Locking`: http://dev.mysql.com/doc/refman/5.1/en/innodb-transaction-model.html
  171. .. _faq-worker-hanging:
  172. celeryd is not doing anything, just hanging
  173. --------------------------------------------
  174. **Answer:** See `MySQL is throwing deadlock errors, what can I do?`_.
  175. or `Why is Task.delay/apply\* just hanging?`.
  176. .. _faq-results-unreliable:
  177. Task results aren't reliably returning
  178. --------------------------------------
  179. **Answer:** If you're using the database backend for results, and in particular
  180. using MySQL, see `MySQL is throwing deadlock errors, what can I do?`_.
  181. .. _faq-publish-hanging:
  182. Why is Task.delay/apply\*/celeryd just hanging?
  183. -----------------------------------------------
  184. **Answer:** There is a bug in some AMQP clients that will make it hang if
  185. it's not able to authenticate the current user, the password doesn't match or
  186. the user does not have access to the virtual host specified. Be sure to check
  187. your broker logs (for RabbitMQ that is :file:`/var/log/rabbitmq/rabbit.log` on
  188. most systems), it usually contains a message describing the reason.
  189. .. _faq-celeryd-on-freebsd:
  190. Does it work on FreeBSD?
  191. ------------------------
  192. **Answer:** The multiprocessing pool requires a working POSIX semaphore
  193. implementation which isn't enabled in FreeBSD by default. You have to enable
  194. POSIX semaphores in the kernel and manually recompile multiprocessing.
  195. Luckily, Viktor Petersson has written a tutorial to get you started with
  196. Celery on FreeBSD here:
  197. http://www.playingwithwire.com/2009/10/how-to-get-celeryd-to-work-on-freebsd/
  198. .. _faq-duplicate-key-errors:
  199. I'm having `IntegrityError: Duplicate Key` errors. Why?
  200. ---------------------------------------------------------
  201. **Answer:** See `MySQL is throwing deadlock errors, what can I do?`_.
  202. Thanks to howsthedotcom.
  203. .. _faq-worker-stops-processing:
  204. Why aren't my tasks processed?
  205. ------------------------------
  206. **Answer:** With RabbitMQ you can see how many consumers are currently
  207. receiving tasks by running the following command:
  208. .. code-block:: bash
  209. $ rabbitmqctl list_queues -p <myvhost> name messages consumers
  210. Listing queues ...
  211. celery 2891 2
  212. This shows that there's 2891 messages waiting to be processed in the task
  213. queue, and there are two consumers processing them.
  214. One reason that the queue is never emptied could be that you have a stale
  215. worker process taking the messages hostage. This could happen if celeryd
  216. wasn't properly shut down.
  217. When a message is received by a worker the broker waits for it to be
  218. acknowledged before marking the message as processed. The broker will not
  219. re-send that message to another consumer until the consumer is shut down
  220. properly.
  221. If you hit this problem you have to kill all workers manually and restart
  222. them::
  223. ps auxww | grep celeryd | awk '{print $2}' | xargs kill
  224. You might have to wait a while until all workers have finished the work they're
  225. doing. If it's still hanging after a long time you can kill them by force
  226. with::
  227. ps auxww | grep celeryd | awk '{print $2}' | xargs kill -9
  228. .. _faq-task-does-not-run:
  229. Why won't my Task run?
  230. ----------------------
  231. **Answer:** There might be syntax errors preventing the tasks module being imported.
  232. You can find out if Celery is able to run the task by executing the
  233. task manually:
  234. >>> from myapp.tasks import MyPeriodicTask
  235. >>> MyPeriodicTask.delay()
  236. Watch celeryd`s log file to see if it's able to find the task, or if some
  237. other error is happening.
  238. .. _faq-periodic-task-does-not-run:
  239. Why won't my periodic task run?
  240. -------------------------------
  241. **Answer:** See `Why won't my Task run?`_.
  242. .. _faq-purge-the-queue:
  243. How do I purge all waiting tasks?
  244. ---------------------------------
  245. **Answer:** You can use the ``celery purge`` command to purge
  246. all configured task queues:
  247. .. code-block:: bash
  248. $ celery purge
  249. or programatically:
  250. .. code-block:: python
  251. >>> from celery import current_app as celery
  252. >>> celery.control.purge()
  253. 1753
  254. If you only want to purge messages from a specific queue
  255. you have to use the AMQP API or the :program:`celery amqp` utility:
  256. .. code-block:: bash
  257. $ celery amqp queue.purge <queue name>
  258. The number 1753 is the number of messages deleted.
  259. You can also start :mod:`~celery.bin.celeryd` with the
  260. :option:`--purge` argument, to purge messages when the worker starts.
  261. .. _faq-messages-left-after-purge:
  262. I've purged messages, but there are still messages left in the queue?
  263. ---------------------------------------------------------------------
  264. **Answer:** Tasks are acknowledged (removed from the queue) as soon
  265. as they are actually executed. After the worker has received a task, it will
  266. take some time until it is actually executed, especially if there are a lot
  267. of tasks already waiting for execution. Messages that are not acknowledged are
  268. held on to by the worker until it closes the connection to the broker (AMQP
  269. server). When that connection is closed (e.g. because the worker was stopped)
  270. the tasks will be re-sent by the broker to the next available worker (or the
  271. same worker when it has been restarted), so to properly purge the queue of
  272. waiting tasks you have to stop all the workers, and then purge the tasks
  273. using :func:`celery.control.purge`.
  274. .. _faq-results:
  275. Results
  276. =======
  277. .. _faq-get-result-by-task-id:
  278. How do I get the result of a task if I have the ID that points there?
  279. ----------------------------------------------------------------------
  280. **Answer**: Use `Task.AsyncResult`::
  281. >>> result = MyTask.AsyncResult(task_id)
  282. >>> result.get()
  283. This will give you a :class:`~celery.result.BaseAsyncResult` instance
  284. using the tasks current result backend.
  285. If you need to specify a custom result backend you should use
  286. :class:`celery.result.BaseAsyncResult` directly::
  287. >>> from celery.result import BaseAsyncResult
  288. >>> result = BaseAsyncResult(task_id, backend=...)
  289. >>> result.get()
  290. .. _faq-security:
  291. Security
  292. ========
  293. Isn't using `pickle` a security concern?
  294. ----------------------------------------
  295. **Answer**: Yes, indeed it is.
  296. You are right to have a security concern, as this can indeed be a real issue.
  297. It is essential that you protect against unauthorized
  298. access to your broker, databases and other services transmitting pickled
  299. data.
  300. For the task messages you can set the :setting:`CELERY_TASK_SERIALIZER`
  301. setting to "json" or "yaml" instead of pickle. There is
  302. currently no alternative solution for task results (but writing a
  303. custom result backend using JSON is a simple task)
  304. Note that this is not just something you should be aware of with Celery, for
  305. example also Django uses pickle for its cache client.
  306. Can messages be encrypted?
  307. --------------------------
  308. **Answer**: Some AMQP brokers supports using SSL (including RabbitMQ).
  309. You can enable this using the :setting:`BROKER_USE_SSL` setting.
  310. It is also possible to add additional encryption and security to messages,
  311. if you have a need for this then you should contact the :ref:`mailing-list`.
  312. Is it safe to run :program:`celeryd` as root?
  313. ---------------------------------------------
  314. **Answer**: No!
  315. We're not currently aware of any security issues, but it would
  316. be incredibly naive to assume that they don't exist, so running
  317. the Celery services (:program:`celeryd`, :program:`celerybeat`,
  318. :program:`celeryev`, etc) as an unprivileged user is recommended.
  319. .. _faq-brokers:
  320. Brokers
  321. =======
  322. Why is RabbitMQ crashing?
  323. -------------------------
  324. **Answer:** RabbitMQ will crash if it runs out of memory. This will be fixed in a
  325. future release of RabbitMQ. please refer to the RabbitMQ FAQ:
  326. http://www.rabbitmq.com/faq.html#node-runs-out-of-memory
  327. .. note::
  328. This is no longer the case, RabbitMQ versions 2.0 and above
  329. includes a new persister, that is tolerant to out of memory
  330. errors. RabbitMQ 2.1 or higher is recommended for Celery.
  331. If you're still running an older version of RabbitMQ and experience
  332. crashes, then please upgrade!
  333. Misconfiguration of Celery can eventually lead to a crash
  334. on older version of RabbitMQ. Even if it doesn't crash, this
  335. can still consume a lot of resources, so it is very
  336. important that you are aware of the common pitfalls.
  337. * Events.
  338. Running :mod:`~celery.bin.celeryd` with the :option:`-E`/:option:`--events`
  339. option will send messages for events happening inside of the worker.
  340. Events should only be enabled if you have an active monitor consuming them,
  341. or if you purge the event queue periodically.
  342. * AMQP backend results.
  343. When running with the AMQP result backend, every task result will be sent
  344. as a message. If you don't collect these results, they will build up and
  345. RabbitMQ will eventually run out of memory.
  346. Results expire after 1 day by default. It may be a good idea
  347. to lower this value by configuring the :setting:`CELERY_TASK_RESULT_EXPIRES`
  348. setting.
  349. If you don't use the results for a task, make sure you set the
  350. `ignore_result` option:
  351. .. code-block python
  352. @celery.task(ignore_result=True)
  353. def mytask():
  354. ...
  355. class MyTask(Task):
  356. ignore_result = True
  357. .. _faq-use-celery-with-stomp:
  358. Can I use Celery with ActiveMQ/STOMP?
  359. -------------------------------------
  360. **Answer**: No. It used to be supported by Carrot,
  361. but is not currently supported in Kombu.
  362. .. _faq-non-amqp-missing-features:
  363. What features are not supported when not using an AMQP broker?
  364. --------------------------------------------------------------
  365. This is an incomplete list of features not available when
  366. using the virtual transports:
  367. * Remote control commands (supported only by Redis).
  368. * Monitoring with events may not work in all virtual transports.
  369. * The `header` and `fanout` exchange types
  370. (`fanout` is supported by Redis).
  371. .. _faq-tasks:
  372. Tasks
  373. =====
  374. .. _faq-tasks-connection-reuse:
  375. How can I reuse the same connection when calling tasks?
  376. -------------------------------------------------------
  377. **Answer**: See the :setting:`BROKER_POOL_LIMIT` setting.
  378. The connection pool is enabled by default since version 2.5.
  379. .. _faq-sudo-subprocess:
  380. Sudo in a :mod:`subprocess` returns :const:`None`
  381. -------------------------------------------------
  382. There is a sudo configuration option that makes it illegal for process
  383. without a tty to run sudo::
  384. Defaults requiretty
  385. If you have this configuration in your :file:`/etc/sudoers` file then
  386. tasks will not be able to call sudo when celeryd is running as a daemon.
  387. If you want to enable that, then you need to remove the line from sudoers.
  388. See: http://timelordz.com/wiki/Apache_Sudo_Commands
  389. .. _faq-deletes-unknown-tasks:
  390. Why do workers delete tasks from the queue if they are unable to process them?
  391. ------------------------------------------------------------------------------
  392. **Answer**:
  393. The worker rejects unknown tasks, messages with encoding errors and messages
  394. that doesn't contain the proper fields (as per the task message protocol).
  395. If it did not reject them they could be redelivered again and again,
  396. causing a loop.
  397. Recent versions of RabbitMQ has the ability to configure a dead-letter
  398. queue for exchange, so that rejected messages is moved there.
  399. .. _faq-execute-task-by-name:
  400. Can I call a task by name?
  401. -----------------------------
  402. **Answer**: Yes. Use :func:`celery.execute.send_task`.
  403. You can also call a task by name from any language
  404. that has an AMQP client.
  405. >>> from celery.execute import send_task
  406. >>> send_task("tasks.add", args=[2, 2], kwargs={})
  407. <AsyncResult: 373550e8-b9a0-4666-bc61-ace01fa4f91d>
  408. .. _faq-get-current-task-id:
  409. How can I get the task id of the current task?
  410. ----------------------------------------------
  411. **Answer**: The current id and more is available in the task request::
  412. @celery.task
  413. def mytask():
  414. cache.set(mytask.request.id, "Running")
  415. For more information see :ref:`task-request-info`.
  416. .. _faq-custom-task-ids:
  417. Can I specify a custom task_id?
  418. -------------------------------
  419. **Answer**: Yes. Use the `task_id` argument to :meth:`Task.apply_async`::
  420. >>> task.apply_async(args, kwargs, task_id="...")
  421. Can I use decorators with tasks?
  422. --------------------------------
  423. **Answer**: Yes. But please see note in the sidebar at :ref:`task-basics`.
  424. .. _faq-natural-task-ids:
  425. Can I use natural task ids?
  426. ---------------------------
  427. **Answer**: Yes, but make sure it is unique, as the behavior
  428. for two tasks existing with the same id is undefined.
  429. The world will probably not explode, but at the worst
  430. they can overwrite each others results.
  431. .. _faq-task-callbacks:
  432. How can I run a task once another task has finished?
  433. ----------------------------------------------------
  434. **Answer**: You can safely launch a task inside a task.
  435. Also, a common pattern is to add callbacks to tasks:
  436. .. code-block:: python
  437. from celery.utils.log import get_task_logger
  438. logger = get_task_logger(__name__)
  439. @celery.task
  440. def add(x, y):
  441. return x + y
  442. @celery.task(ignore_result=True)
  443. def log_result(result):
  444. logger.info("log_result got: %r", result)
  445. Invocation::
  446. >>> (add.s(2, 2) | log_result.s()).delay()
  447. See :doc:`userguide/canvas` for more information.
  448. .. _faq-cancel-task:
  449. Can I cancel the execution of a task?
  450. -------------------------------------
  451. **Answer**: Yes. Use `result.revoke`::
  452. >>> result = add.apply_async(args=[2, 2], countdown=120)
  453. >>> result.revoke()
  454. or if you only have the task id::
  455. >>> from celery import current_app as celery
  456. >>> celery.control.revoke(task_id)
  457. .. _faq-node-not-receiving-broadcast-commands:
  458. Why aren't my remote control commands received by all workers?
  459. --------------------------------------------------------------
  460. **Answer**: To receive broadcast remote control commands, every worker node
  461. uses its host name to create a unique queue name to listen to,
  462. so if you have more than one worker with the same host name, the
  463. control commands will be received in round-robin between them.
  464. To work around this you can explicitly set the host name for every worker
  465. using the :option:`--hostname` argument to :mod:`~celery.bin.celeryd`:
  466. .. code-block:: bash
  467. $ celeryd --hostname=$(hostname).1
  468. $ celeryd --hostname=$(hostname).2
  469. etc., etc...
  470. .. _faq-task-routing:
  471. Can I send some tasks to only some servers?
  472. --------------------------------------------
  473. **Answer:** Yes. You can route tasks to an arbitrary server using AMQP,
  474. and a worker can bind to as many queues as it wants.
  475. See :doc:`userguide/routing` for more information.
  476. .. _faq-change-periodic-task-interval-at-runtime:
  477. Can I change the interval of a periodic task at runtime?
  478. --------------------------------------------------------
  479. **Answer**: Yes. You can use the Django database scheduler, or you can
  480. override `PeriodicTask.is_due` or turn `PeriodicTask.run_every` into a
  481. property:
  482. .. code-block:: python
  483. class MyPeriodic(PeriodicTask):
  484. def run(self):
  485. # ...
  486. @property
  487. def run_every(self):
  488. return get_interval_from_database(...)
  489. .. _faq-task-priorities:
  490. Does celery support task priorities?
  491. ------------------------------------
  492. **Answer**: No. In theory, yes, as AMQP supports priorities. However
  493. RabbitMQ doesn't implement them yet.
  494. The usual way to prioritize work in Celery, is to route high priority tasks
  495. to different servers. In the real world this may actually work better than per message
  496. priorities. You can use this in combination with rate limiting to achieve a
  497. highly responsive system.
  498. .. _faq-acks_late-vs-retry:
  499. Should I use retry or acks_late?
  500. --------------------------------
  501. **Answer**: Depends. It's not necessarily one or the other, you may want
  502. to use both.
  503. `Task.retry` is used to retry tasks, notably for expected errors that
  504. is catchable with the `try:` block. The AMQP transaction is not used
  505. for these errors: **if the task raises an exception it is still acknowledged!**.
  506. The `acks_late` setting would be used when you need the task to be
  507. executed again if the worker (for some reason) crashes mid-execution.
  508. It's important to note that the worker is not known to crash, and if
  509. it does it is usually an unrecoverable error that requires human
  510. intervention (bug in the worker, or task code).
  511. In an ideal world you could safely retry any task that has failed, but
  512. this is rarely the case. Imagine the following task:
  513. .. code-block:: python
  514. @celery.task
  515. def process_upload(filename, tmpfile):
  516. # Increment a file count stored in a database
  517. increment_file_counter()
  518. add_file_metadata_to_db(filename, tmpfile)
  519. copy_file_to_destination(filename, tmpfile)
  520. If this crashed in the middle of copying the file to its destination
  521. the world would contain incomplete state. This is not a critical
  522. scenario of course, but you can probably imagine something far more
  523. sinister. So for ease of programming we have less reliability;
  524. It's a good default, users who require it and know what they
  525. are doing can still enable acks_late (and in the future hopefully
  526. use manual acknowledgement)
  527. In addition `Task.retry` has features not available in AMQP
  528. transactions: delay between retries, max retries, etc.
  529. So use retry for Python errors, and if your task is idempotent
  530. combine that with `acks_late` if that level of reliability
  531. is required.
  532. .. _faq-schedule-at-specific-time:
  533. Can I schedule tasks to execute at a specific time?
  534. ---------------------------------------------------
  535. .. module:: celery.task.base
  536. **Answer**: Yes. You can use the `eta` argument of :meth:`Task.apply_async`.
  537. Or to schedule a periodic task at a specific time, use the
  538. :class:`celery.schedules.crontab` schedule behavior:
  539. .. code-block:: python
  540. from celery.schedules import crontab
  541. from celery.task import periodic_task
  542. @periodic_task(run_every=crontab(hour=7, minute=30, day_of_week="mon"))
  543. def every_monday_morning():
  544. print("This is run every Monday morning at 7:30")
  545. .. _faq-safe-worker-shutdown:
  546. How do I shut down `celeryd` safely?
  547. --------------------------------------
  548. **Answer**: Use the :sig:`TERM` signal, and the worker will finish all currently
  549. executing jobs and shut down as soon as possible. No tasks should be lost.
  550. You should never stop :mod:`~celery.bin.celeryd` with the :sig:`KILL` signal
  551. (:option:`-9`), unless you've tried :sig:`TERM` a few times and waited a few
  552. minutes to let it get a chance to shut down. As if you do tasks may be
  553. terminated mid-execution, and they will not be re-run unless you have the
  554. `acks_late` option set (`Task.acks_late` / :setting:`CELERY_ACKS_LATE`).
  555. .. seealso::
  556. :ref:`worker-stopping`
  557. .. _faq-daemonizing:
  558. How do I run celeryd in the background on [platform]?
  559. -----------------------------------------------------
  560. **Answer**: Please see :ref:`daemonizing`.
  561. .. _faq-django:
  562. Django
  563. ======
  564. .. _faq-django-database-tables:
  565. What purpose does the database tables created by django-celery have?
  566. --------------------------------------------------------------------
  567. Several database tables are created by default, these relate to
  568. * Monitoring
  569. When you use the django-admin monitor, the cluster state is written
  570. to the ``TaskState`` and ``WorkerState`` models.
  571. * Periodic tasks
  572. When the database-backed schedule is used the periodic task
  573. schedule is taken from the ``PeriodicTask`` model, there are
  574. also several other helper tables (``IntervalSchedule``,
  575. ``CrontabSchedule``, ``PeriodicTasks``).
  576. * Task results
  577. The database result backend is enabled by default when using django-celery
  578. (this is for historical reasons, and thus for backward compatibility).
  579. The results are stored in the ``TaskMeta`` and ``TaskSetMeta`` models.
  580. *these tables are not created if another result backend is configured*.
  581. .. _faq-windows:
  582. Windows
  583. =======
  584. .. _faq-windows-worker-spawn-loop:
  585. celeryd keeps spawning processes at startup
  586. -------------------------------------------
  587. **Answer**: This is a known issue on Windows.
  588. You have to start celeryd with the command:
  589. .. code-block:: bash
  590. $ python -m celery.bin.celeryd
  591. Any additional arguments can be appended to this command.
  592. See http://bit.ly/bo9RSw
  593. .. _faq-windows-worker-embedded-beat:
  594. The `-B` / `--beat` option to celeryd doesn't work?
  595. ----------------------------------------------------------------
  596. **Answer**: That's right. Run `celerybeat` and `celeryd` as separate
  597. services instead.