Changelog 74 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762176317641765176617671768176917701771177217731774177517761777177817791780178117821783178417851786178717881789179017911792179317941795179617971798179918001801180218031804180518061807180818091810181118121813181418151816181718181819182018211822182318241825182618271828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885188618871888188918901891189218931894189518961897189818991900190119021903190419051906190719081909191019111912191319141915191619171918191919201921192219231924192519261927192819291930193119321933193419351936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005200620072008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920402041204220432044204520462047204820492050205120522053205420552056205720582059206020612062206320642065206620672068206920702071207220732074207520762077207820792080208120822083208420852086208720882089209020912092209320942095209620972098209921002101210221032104210521062107210821092110211121122113211421152116211721182119212021212122212321242125212621272128212921302131213221332134213521362137213821392140214121422143214421452146214721482149215021512152215321542155215621572158215921602161216221632164216521662167216821692170217121722173217421752176217721782179218021812182218321842185218621872188218921902191219221932194219521962197219821992200220122022203220422052206220722082209221022112212221322142215221622172218221922202221222222232224222522262227222822292230223122322233223422352236223722382239224022412242224322442245224622472248224922502251225222532254225522562257
  1. ================
  2. Change history
  3. ================
  4. .. contents::
  5. :local:
  6. 2.0.0
  7. =====
  8. :release-date: NOT RELEASED
  9. :branch: master
  10. :state: freeze
  11. Foreword
  12. --------
  13. Celery 2.0 contains backward incompatible changes, the most important
  14. being that the Django dependency has been removed so Celery no longer
  15. supports Django out of the box, but instead as an add-on package
  16. called `django-celery`_.
  17. We're very sorry for breaking backwards compatibility, but there's
  18. also many new and exciting features to make up for the time you lose
  19. upgrading, so be sure to read the :ref:`News <120news>` section.
  20. Quite a lot of potential users have been upset about the Django dependency,
  21. so maybe this is a chance to get wider adoption by the Python community as
  22. well.
  23. Big thanks to all contributors, testers and users!
  24. Upgrading for Django-users
  25. --------------------------
  26. Django integration has been moved to a separate package: `django-celery`_.
  27. * To upgrade you need to install the `django-celery`_ module and change::
  28. INSTALLED_APPS = "celery"
  29. to::
  30. INSTALLED_APPS = "djcelery"
  31. * If you use ``mod_wsgi`` you need to add the following line to your ``.wsgi``
  32. file::
  33. import os
  34. os.environ["CELERY_LOADER"] = "django"
  35. * The following modules has been moved to `django-celery`_:
  36. ===================================== =====================================
  37. **Module name** **Replace with**
  38. ===================================== =====================================
  39. ``celery.models`` ``djcelery.models``
  40. ``celery.managers`` ``djcelery.managers``
  41. ``celery.views`` ``djcelery.views``
  42. ``celery.urls`` ``djcelery.urls``
  43. ``celery.management`` ``djcelery.management``
  44. ``celery.loaders.djangoapp`` ``djcelery.loaders``
  45. ``celery.backends.database`` ``djcelery.backends.database``
  46. ``celery.backends.cache`` ``djcelery.backends.cache``
  47. ===================================== =====================================
  48. Importing :mod:`djcelery` will automatically setup Celery to use the Django.
  49. It does this by pointing the :envvar:`CELERY_LOADER` environment variable to
  50. the Django loader (it won't change it if a loader is already set.)
  51. When the Django loader is used, the "database" and "cache" result backend
  52. aliases will point to the :mod:`djcelery` backends instead of the built-in backends,
  53. and configuration will be read from the Django settings.
  54. .. _`django-celery`: http://pypi.python.org/pypi/django-celery
  55. Upgrading for others
  56. --------------------
  57. Database result backend
  58. ~~~~~~~~~~~~~~~~~~~~~~~
  59. The database result backend is now using `SQLAlchemy`_ instead of the
  60. Django ORM, see `Supported Databases`_ for a table of supported databases.
  61. The ``DATABASE_*`` settings has been replaced by a single setting:
  62. ``CELERY_RESULT_DBURI``. The value here should be an
  63. `SQLAlchemy Connection String`_, some examples include:
  64. .. code-block:: python
  65. # sqlite (filename)
  66. CELERY_RESULT_DBURI = "sqlite:///celerydb.sqlite"
  67. # mysql
  68. CELERY_RESULT_DBURI = "mysql://scott:tiger@localhost/foo"
  69. # postgresql
  70. CELERY_RESULT_DBURI = "postgresql://scott:tiger@localhost/mydatabase"
  71. # oracle
  72. CELERY_RESULT_DBURI = "oracle://scott:tiger@127.0.0.1:1521/sidname"
  73. See `SQLAlchemy Connection Strings`_ for more information about connection
  74. strings.
  75. To specify additional SQLAlchemy database engine options you can use
  76. the ``CELERY_RESULT_ENGINE_OPTIONS`` setting::
  77. # echo enables verbose logging from SQLAlchemy.
  78. CELERY_RESULT_ENGINE_OPTIONS = {"echo": True}
  79. .. _`SQLAlchemy`:
  80. http://www.sqlalchemy.org
  81. .. _`Supported Databases`:
  82. http://www.sqlalchemy.org/docs/dbengine.html#supported-databases
  83. .. _`SQLAlchemy Connection String`:
  84. http://www.sqlalchemy.org/docs/dbengine.html#create-engine-url-arguments
  85. .. _`SQLAlchemy Connection Strings`:
  86. http://www.sqlalchemy.org/docs/dbengine.html#create-engine-url-arguments
  87. Cache result backend
  88. ~~~~~~~~~~~~~~~~~~~~
  89. The cache result backend is no longer using the Django cache framework,
  90. but it supports mostly the same configuration syntax::
  91. CELERY_CACHE_BACKEND = "memcached://A.example.com:11211;B.example.com"
  92. To use the cache backend you must either have the `pylibmc`_ or
  93. `python-memcached`_ library installed, of which the former is regarded
  94. as the best choice.
  95. .. _`pylibmc`: http://pypi.python.org/pypi/pylibmc
  96. .. _`python-memcached`: http://pypi.python.org/pypi/python-memcached
  97. The support backend types are ``memcached://`` and ``memory://``,
  98. we haven't felt the need to support any of the other backends
  99. provided by Django.
  100. Backward incompatible changes
  101. -----------------------------
  102. * Default (python) loader now prints warning on missing ``celeryconfig.py``
  103. instead of raising :exc:`ImportError`.
  104. celeryd raises :exc:`~celery.exceptions.ImproperlyConfigured` if the configuration
  105. is not set up. This makes it possible to use ``--help`` etc, without having a
  106. working configuration.
  107. Also this makes it possible to use the client side of celery without being
  108. configured::
  109. >>> from carrot.connection import BrokerConnection
  110. >>> conn = BrokerConnection("localhost", "guest", "guest", "/")
  111. >>> from celery.execute import send_task
  112. >>> r = send_task("celery.ping", args=(), kwargs={}, connection=conn)
  113. >>> from celery.backends.amqp import AMQPBackend
  114. >>> r.backend = AMQPBackend(connection=conn)
  115. >>> r.get()
  116. 'pong'
  117. * The following deprecated settings has been removed (as scheduled by
  118. the `deprecation timeline`_):
  119. ===================================== =====================================
  120. **Setting name** **Replace with**
  121. ===================================== =====================================
  122. ``CELERY_AMQP_CONSUMER_QUEUES`` ``CELERY_QUEUES``
  123. ``CELERY_AMQP_EXCHANGE`` ``CELERY_DEFAULT_EXCHANGE``
  124. ``CELERY_AMQP_EXCHANGE_TYPE`` ``CELERY_DEFAULT_EXCHANGE_TYPE``
  125. ``CELERY_AMQP_CONSUMER_ROUTING_KEY`` ``CELERY_QUEUES``
  126. ``CELERY_AMQP_PUBLISHER_ROUTING_KEY`` ``CELERY_DEFAULT_ROUTING_KEY``
  127. ===================================== =====================================
  128. .. _`deprecation timeline`:
  129. http://ask.github.com/celery/internals/deprecation.html
  130. * The ``celery.task.rest`` module has been removed, use :mod:`celery.task.http`
  131. instead (as scheduled by the `deprecation timeline`_).
  132. * It's no longer allowed to skip the class name in loader names.
  133. (as scheduled by the `deprecation timeline`_):
  134. Assuming the implicit ``Loader`` class name is no longer supported,
  135. if you use e.g.::
  136. CELERY_LOADER = "myapp.loaders"
  137. You need to include the loader class name, like this::
  138. CELERY_LOADER = "myapp.loaders.Loader"
  139. * ``CELERY_TASK_RESULT_EXPIRES`` now defaults to 1 day.
  140. Previous default setting was to expire in 5 days.
  141. * AMQP backend: Don't use different values for `auto_delete`.
  142. This bug became visible with RabbitMQ 1.8.0, which no longer
  143. allows conflicting declarations for the auto_delete and durable settings.
  144. If you've already used celery with this backend chances are you
  145. have to delete the previous declaration::
  146. $ PYTHONPATH=. camqadm exchange.delete celeryresults
  147. * Now uses pickle instead of cPickle on Python versions <= 2.5
  148. cPikle is broken in Python <= 2.5.
  149. It unsafely and incorrectly uses relative instead of absolute imports,
  150. so e.g::
  151. exceptions.KeyError
  152. becomes::
  153. celery.exceptions.KeyError
  154. Your best choice is to upgrade to Python 2.6,
  155. as while the pure pickle version has worse performance,
  156. it is the only safe option for older Python versions.
  157. .. _120news:
  158. News
  159. ----
  160. * **celeryev**: Curses Celery Monitor and Event Viewer.
  161. This is a simple monitor allowing you to see what tasks are
  162. executing in real-time and investigate tracebacks and results of ready
  163. tasks. It also enables you to set new rate limits and revoke tasks.
  164. Screenshot:
  165. .. image:: http://celeryproject.org/img/celeryevshotsm.jpg
  166. If you run ``celeryev`` with the ``-d`` switch it will act as an event
  167. dumper, simply dumping the events it receives to standard out::
  168. $ celeryev -d
  169. -> celeryev: starting capture...
  170. casper.local [2010-06-04 10:42:07.020000] heartbeat
  171. casper.local [2010-06-04 10:42:14.750000] task received:
  172. tasks.add(61a68756-27f4-4879-b816-3cf815672b0e) args=[2, 2] kwargs={}
  173. eta=2010-06-04T10:42:16.669290, retries=0
  174. casper.local [2010-06-04 10:42:17.230000] task started
  175. tasks.add(61a68756-27f4-4879-b816-3cf815672b0e) args=[2, 2] kwargs={}
  176. casper.local [2010-06-04 10:42:17.960000] task succeeded:
  177. tasks.add(61a68756-27f4-4879-b816-3cf815672b0e)
  178. args=[2, 2] kwargs={} result=4, runtime=0.782663106918
  179. The fields here are, in order: *sender hostname*, *timestamp*, *event type* and
  180. *additional event fields*.
  181. * AMQP result backend: Now supports ``.ready()``, ``.successful()``,
  182. ``.result``, ``.status``, and even responds to changes in task state
  183. * New user guides:
  184. * :doc:`userguide/workers`
  185. * :doc:`userguide/tasksets`
  186. * :doc:`userguide/routing`
  187. * celeryd: Standard out/error is now being redirected to the logfile.
  188. * :mod:`billiard` has been moved back to the celery repository.
  189. ===================================== =====================================
  190. **Module name** **celery equivalent**
  191. ===================================== =====================================
  192. ``billiard.pool`` ``celery.concurrency.processes.pool``
  193. ``billiard.serialization`` ``celery.serialization``
  194. ``billiard.utils.functional`` ``celery.utils.functional``
  195. ===================================== =====================================
  196. The :mod:`billiard` distribution may be maintained, depending on interest.
  197. * now depends on :mod:`carrot` >= 0.10.5
  198. * now depends on :mod:`pyparsing`
  199. * celeryd: Added ``--purge`` as an alias to ``--discard``.
  200. * celeryd: Ctrl+C (SIGINT) once does warm shutdown, hitting Ctrl+C twice
  201. forces termination.
  202. * Added support for using complex crontab-expressions in periodic tasks. For
  203. example, you can now use::
  204. >>> crontab(minute="*/15")
  205. or even::
  206. >>> crontab(minute="*/30", hour="8-17,1-2", day_of_week="thu-fri")
  207. See :doc:`getting-started/periodic-tasks`.
  208. * celeryd: Now waits for available pool processes before applying new
  209. tasks to the pool.
  210. This means it doesn't have to wait for dozens of tasks to finish at shutdown
  211. because it has applied prefetched tasks without having any pool
  212. processes available to immediately accept them.
  213. See http://github.com/ask/celery/issues/closed#issue/122
  214. * New built-in way to do task callbacks using
  215. :class:`~celery.task.sets.subtask`.
  216. See :doc:`userguide/tasksets` for more information.
  217. * TaskSets can now contain several types of tasks.
  218. :class:`~celery.task.sets.TaskSet` has been refactored to use
  219. a new syntax, please see :doc:`userguide/tasksets` for more information.
  220. The previous syntax is still supported, but will be deprecated in
  221. version 1.4.
  222. * TaskSet failed() result was incorrect.
  223. See http://github.com/ask/celery/issues/closed#issue/132
  224. * Now creates different loggers per task class.
  225. See http://github.com/ask/celery/issues/closed#issue/129
  226. * Missing queue definitions are now created automatically.
  227. You can disable this using the CELERY_CREATE_MISSING_QUEUES setting.
  228. The missing queues are created with the following options::
  229. CELERY_QUEUES[name] = {"exchange": name,
  230. "exchange_type": "direct",
  231. "routing_key": "name}
  232. This feature is added for easily setting up routing using the ``-Q``
  233. option to ``celeryd``::
  234. $ celeryd -Q video, image
  235. See the new routing section of the userguide for more information:
  236. :doc:`userguide/routing`.
  237. * New Task option: ``Task.queue``
  238. If set, message options will be taken from the corresponding entry
  239. in ``CELERY_QUEUES``. ``exchange``, ``exchange_type`` and ``routing_key``
  240. will be ignored
  241. * Added support for task soft and hard timelimits.
  242. New settings added:
  243. * CELERYD_TASK_TIME_LIMIT
  244. Hard time limit. The worker processing the task will be killed and
  245. replaced with a new one when this is exceeded.
  246. * CELERYD_SOFT_TASK_TIME_LIMIT
  247. Soft time limit. The celery.exceptions.SoftTimeLimitExceeded exception
  248. will be raised when this is exceeded. The task can catch this to
  249. e.g. clean up before the hard time limit comes.
  250. New command line arguments to celeryd added:
  251. ``--time-limit`` and ``--soft-time-limit``.
  252. What's left?
  253. This won't work on platforms not supporting signals (and specifically
  254. the ``SIGUSR1`` signal) yet. So an alternative the ability to disable
  255. the feature alltogether on nonconforming platforms must be implemented.
  256. Also when the hard time limit is exceeded, the task result should
  257. be a ``TimeLimitExceeded`` exception.
  258. * Test suite is now passing without a running broker, using the carrot
  259. in-memory backend.
  260. * Log output is now available in colors.
  261. ===================================== =====================================
  262. **Log level** **Color**
  263. ===================================== =====================================
  264. ``DEBUG`` Blue
  265. ``WARNING`` Yellow
  266. ``CRITICAL`` Magenta
  267. ``ERROR`` Red
  268. ===================================== =====================================
  269. This is only enabled when the log output is a tty.
  270. You can explicitly enable/disable this feature using the
  271. ``CELERYD_LOG_COLOR`` setting.
  272. * Added support for task router classes (like the django multidb routers)
  273. * New setting: CELERY_ROUTES
  274. This is a single, or a list of routers to traverse when
  275. sending tasks. Dicts in this list converts to a
  276. :class:`celery.routes.MapRoute` instance.
  277. Examples:
  278. >>> CELERY_ROUTES = {"celery.ping": "default",
  279. "mytasks.add": "cpu-bound",
  280. "video.encode": {
  281. "queue": "video",
  282. "exchange": "media"
  283. "routing_key": "media.video.encode"}}
  284. >>> CELERY_ROUTES = ("myapp.tasks.Router",
  285. {"celery.ping": "default})
  286. Where ``myapp.tasks.Router`` could be:
  287. .. code-block:: python
  288. class Router(object):
  289. def route_for_task(self, task, args=None, kwargs=None):
  290. if task == "celery.ping":
  291. return "default"
  292. route_for_task may return a string or a dict. A string then means
  293. it's a queue name in ``CELERY_QUEUES``, a dict means it's a custom route.
  294. When sending tasks, the routers are consulted in order. The first
  295. router that doesn't return ``None`` is the route to use. The message options
  296. is then merged with the found route settings, where the routers settings
  297. have priority.
  298. Example if :func:`~celery.execute.apply_async` has these arguments::
  299. >>> Task.apply_async(immediate=False, exchange="video",
  300. ... routing_key="video.compress")
  301. and a router returns::
  302. {"immediate": True,
  303. "exchange": "urgent"}
  304. the final message options will be::
  305. immediate=True, exchange="urgent", routing_key="video.compress"
  306. (and any default message options defined in the
  307. :class:`~celery.task.base.Task` class)
  308. * New Task handler called after the task returns:
  309. :meth:`~celery.task.base.Task.after_return`.
  310. * :class:`~celery.datastructures.ExceptionInfo` now passed to
  311. :meth:`~celery.task.base.Task.on_retry`/
  312. :meth:`~celery.task.base.Task.on_failure` as einfo keyword argument.
  313. * celeryd: Added ``CELERYD_MAX_TASKS_PER_CHILD`` /
  314. :option:`--maxtasksperchild`
  315. Defines the maximum number of tasks a pool worker can process before
  316. the process is terminated and replaced by a new one.
  317. * Revoked tasks now marked with state ``REVOKED``, and ``result.get()``
  318. will now raise :exc:`~celery.exceptions.TaskRevokedError`.
  319. * :func:`celery.task.control.ping` now works as expected.
  320. * ``apply(throw=True)`` / ``CELERY_EAGER_PROPAGATES_EXCEPTIONS``: Makes eager
  321. execution re-raise task errors.
  322. * New signal: :data:`~celery.signals.worker_process_init`: Sent inside the
  323. pool worker process at init.
  324. * celeryd :option:`-Q` option: Ability to specifiy list of queues to use,
  325. disabling other configured queues.
  326. For example, if ``CELERY_QUEUES`` defines four queues: ``image``, ``video``,
  327. ``data`` and ``default``, the following command would make celeryd only
  328. consume from the ``image`` and ``video`` queues::
  329. $ celeryd -Q image,video
  330. * celeryd: New return value for the ``revoke`` control command:
  331. Now returns::
  332. {"ok": "task $id revoked"}
  333. instead of ``True``.
  334. * celeryd: Can now enable/disable events using remote control
  335. Example usage:
  336. >>> from celery.task.control import broadcast
  337. >>> broadcast("enable_events")
  338. >>> broadcast("disable_events")
  339. * Removed top-level tests directory. Test config now in celery.tests.config
  340. This means running the unittests doesn't require any special setup.
  341. ``celery/tests/__init__`` now configures the ``CELERY_CONFIG_MODULE`` and
  342. ``CELERY_LOADER``, so when ``nosetests`` imports that, the unit test
  343. environment is all set up.
  344. Before you run the tests you need to install the test requirements::
  345. $ pip install -r contrib/requirements/test.txt
  346. Running all tests::
  347. $ nosetests
  348. Specifying the tests to run::
  349. $ nosetests celery.tests.test_task
  350. Producing HTML coverage::
  351. $ nosetests --with-coverage3
  352. The coverage output is then located in ``celery/tests/cover/index.html``.
  353. * celeryd: New option ``--version``: Dump version info and exit.
  354. * :mod:`celeryd-multi <celeryd.bin.celeryd_multi>`: Tool for shell scripts
  355. to start multiple workers.
  356. Some examples::
  357. # Advanced example with 10 workers:
  358. # * Three of the workers processes the images and video queue
  359. # * Two of the workers processes the data queue with loglevel DEBUG
  360. # * the rest processes the default' queue.
  361. $ celeryd-multi start 10 -l INFO -Q:1-3 images,video -Q:4,5:data
  362. -Q default -L:4,5 DEBUG
  363. # get commands to start 10 workers, with 3 processes each
  364. $ celeryd-multi start 3 -c 3
  365. celeryd -n celeryd1.myhost -c 3
  366. celeryd -n celeryd2.myhost -c 3
  367. celeryd- n celeryd3.myhost -c 3
  368. # start 3 named workers
  369. $ celeryd-multi start image video data -c 3
  370. celeryd -n image.myhost -c 3
  371. celeryd -n video.myhost -c 3
  372. celeryd -n data.myhost -c 3
  373. # specify custom hostname
  374. $ celeryd-multi start 2 -n worker.example.com -c 3
  375. celeryd -n celeryd1.worker.example.com -c 3
  376. celeryd -n celeryd2.worker.example.com -c 3
  377. # Additionl options are added to each celeryd',
  378. # but you can also modify the options for ranges of or single workers
  379. # 3 workers: Two with 3 processes, and one with 10 processes.
  380. $ celeryd-multi start 3 -c 3 -c:1 10
  381. celeryd -n celeryd1.myhost -c 10
  382. celeryd -n celeryd2.myhost -c 3
  383. celeryd -n celeryd3.myhost -c 3
  384. # can also specify options for named workers
  385. $ celeryd-multi start image video data -c 3 -c:image 10
  386. celeryd -n image.myhost -c 10
  387. celeryd -n video.myhost -c 3
  388. celeryd -n data.myhost -c 3
  389. # ranges and lists of workers in options is also allowed:
  390. # (-c:1-3 can also be written as -c:1,2,3)
  391. $ celeryd-multi start 5 -c 3 -c:1-3 10
  392. celeryd-multi -n celeryd1.myhost -c 10
  393. celeryd-multi -n celeryd2.myhost -c 10
  394. celeryd-multi -n celeryd3.myhost -c 10
  395. celeryd-multi -n celeryd4.myhost -c 3
  396. celeryd-multi -n celeryd5.myhost -c 3
  397. # lists also works with named workers
  398. $ celeryd-multi start foo bar baz xuzzy -c 3 -c:foo,bar,baz 10
  399. celeryd-multi -n foo.myhost -c 10
  400. celeryd-multi -n bar.myhost -c 10
  401. celeryd-multi -n baz.myhost -c 10
  402. celeryd-multi -n xuzzy.myhost -c 3
  403. * The worker now calls the result backends ``process_cleanup`` method
  404. *after* task execution instead of before.
  405. * AMQP result backend now supports Pika.
  406. 1.0.6
  407. =====
  408. :release-date: 2010-06-30 09:57 A.M CEST
  409. * RabbitMQ 1.8.0 has extended their exchange equivalence tests to
  410. include ``auto_delete`` and ``durable``. This broke the AMQP backend.
  411. If you've already used the AMQP backend this means you have to
  412. delete the previous definitions::
  413. $ PYTHONPATH=. camqadm exchange.delete celeryresults
  414. or::
  415. $ python manage.py camqadm exchange.delete celeryresults
  416. 1.0.5
  417. =====
  418. :release-date: 2010-06-01 02:36 P.M CEST
  419. Critical
  420. --------
  421. * SIGINT/Ctrl+C killed the pool, abrubtly terminating the currently executing
  422. tasks.
  423. Fixed by making the pool worker processes ignore :const:`SIGINT`.
  424. * Should not close the consumers before the pool is terminated, just cancel the consumers.
  425. Issue #122. http://github.com/ask/celery/issues/issue/122
  426. * Now depends on :mod:`billiard` >= 0.3.1
  427. * celeryd: Previously exceptions raised by worker components could stall startup,
  428. now it correctly logs the exceptions and shuts down.
  429. * celeryd: Prefetch counts was set too late. QoS is now set as early as possible,
  430. so celeryd can't slurp in all the messages at start-up.
  431. Changes
  432. -------
  433. * :mod:`celery.contrib.abortable`: Abortable tasks.
  434. Tasks that defines steps of execution, the task can then
  435. be aborted after each step has completed.
  436. * :class:`~celery.events.EventDispatcher`: No longer creates AMQP channel
  437. if events are disabled
  438. * Added required RPM package names under ``[bdist_rpm]`` section, to support building RPMs
  439. from the sources using setup.py
  440. * Running unittests: :envvar:`NOSE_VERBOSE` environment var now enables verbose output from Nose.
  441. * :func:`celery.execute.apply`: Pass logfile/loglevel arguments as task kwargs.
  442. Issue #110 http://github.com/ask/celery/issues/issue/110
  443. * celery.execute.apply: Should return exception, not :class:`~celery.datastructures.ExceptionInfo`
  444. on error.
  445. Issue #111 http://github.com/ask/celery/issues/issue/111
  446. * Added new entries to the :doc:`FAQs <faq>`:
  447. * Should I use retry or acks_late?
  448. * Can I execute a task by name?
  449. 1.0.4
  450. =====
  451. :release-date: 2010-05-31 09:54 A.M CEST
  452. * Changlog merged with 1.0.5 as the release was never announced.
  453. 1.0.3
  454. =====
  455. :release-date: 2010-05-15 03:00 P.M CEST
  456. Important notes
  457. ---------------
  458. * Messages are now acked *just before* the task function is executed.
  459. This is the behavior we've wanted all along, but couldn't have because of
  460. limitations in the multiprocessing module.
  461. The previous behavior was not good, and the situation worsened with the
  462. release of 1.0.1, so this change will definitely improve
  463. reliability, performance and operations in general.
  464. For more information please see http://bit.ly/9hom6T
  465. * Database result backend: result now explicitly sets ``null=True`` as
  466. ``django-picklefield`` version 0.1.5 changed the default behavior
  467. right under our noses :(
  468. See: http://bit.ly/d5OwMr
  469. This means those who created their celery tables (via syncdb or
  470. celeryinit) with picklefield versions >= 0.1.5 has to alter their tables to
  471. allow the result field to be ``NULL`` manually.
  472. MySQL::
  473. ALTER TABLE celery_taskmeta MODIFY result TEXT NULL
  474. PostgreSQL::
  475. ALTER TABLE celery_taskmeta ALTER COLUMN result DROP NOT NULL
  476. * Removed ``Task.rate_limit_queue_type``, as it was not really useful
  477. and made it harder to refactor some parts.
  478. * Now depends on carrot >= 0.10.4
  479. * Now depends on billiard >= 0.3.0
  480. News
  481. ----
  482. * AMQP backend: Added timeout support for ``result.get()`` /
  483. ``result.wait()``.
  484. * New task option: ``Task.acks_late`` (default: ``CELERY_ACKS_LATE``)
  485. Late ack means the task messages will be acknowledged **after** the task
  486. has been executed, not *just before*, which is the default behavior.
  487. Note that this means the tasks may be executed twice if the worker
  488. crashes in the middle of their execution. Not acceptable for most
  489. applications, but desirable for others.
  490. * Added crontab-like scheduling to periodic tasks.
  491. Like a cron job, you can specify units of time of when
  492. you would like the task to execute. While not a full implementation
  493. of cron's features, it should provide a fair degree of common scheduling
  494. needs.
  495. You can specify a minute (0-59), an hour (0-23), and/or a day of the
  496. week (0-6 where 0 is Sunday, or by names: sun, mon, tue, wed, thu, fri,
  497. sat).
  498. Examples:
  499. .. code-block:: python
  500. from celery.task.schedules import crontab
  501. from celery.decorators import periodic_task
  502. @periodic_task(run_every=crontab(hour=7, minute=30))
  503. def every_morning():
  504. print("Runs every morning at 7:30a.m")
  505. @periodic_task(run_every=crontab(hour=7, minute=30, day_of_week="mon"))
  506. def every_monday_morning():
  507. print("Run every monday morning at 7:30a.m")
  508. @periodic_task(run_every=crontab(minutes=30))
  509. def every_hour():
  510. print("Runs every hour on the clock. e.g. 1:30, 2:30, 3:30 etc.")
  511. Note that this a late addition. While we have unittests, due to the
  512. nature of this feature we haven't been able to completely test this
  513. in practice, so consider this experimental.
  514. * ``TaskPool.apply_async``: Now supports the ``accept_callback`` argument.
  515. * ``apply_async``: Now raises :exc:`ValueError` if task args is not a list,
  516. or kwargs is not a tuple (http://github.com/ask/celery/issues/issue/95).
  517. * ``Task.max_retries`` can now be ``None``, which means it will retry forever.
  518. * Celerybeat: Now reuses the same connection when publishing large
  519. sets of tasks.
  520. * Modified the task locking example in the documentation to use
  521. ``cache.add`` for atomic locking.
  522. * Added experimental support for a *started* status on tasks.
  523. If ``Task.track_started`` is enabled the task will report its status
  524. as "started" when the task is executed by a worker.
  525. The default value is ``False`` as the normal behaviour is to not
  526. report that level of granularity. Tasks are either pending, finished,
  527. or waiting to be retried. Having a "started" status can be useful for
  528. when there are long running tasks and there is a need to report which
  529. task is currently running.
  530. The global default can be overridden by the ``CELERY_TRACK_STARTED``
  531. setting.
  532. * User Guide: New section ``Tips and Best Practices``.
  533. Contributions welcome!
  534. Remote control commands
  535. -----------------------
  536. * Remote control commands can now send replies back to the caller.
  537. Existing commands has been improved to send replies, and the client
  538. interface in ``celery.task.control`` has new keyword arguments: ``reply``,
  539. ``timeout`` and ``limit``. Where reply means it will wait for replies,
  540. timeout is the time in seconds to stop waiting for replies, and limit
  541. is the maximum number of replies to get.
  542. By default, it will wait for as many replies as possible for one second.
  543. * rate_limit(task_name, destination=all, reply=False, timeout=1, limit=0)
  544. Worker returns ``{"ok": message}`` on success,
  545. or ``{"failure": message}`` on failure.
  546. >>> from celery.task.control import rate_limit
  547. >>> rate_limit("tasks.add", "10/s", reply=True)
  548. [{'worker1': {'ok': 'new rate limit set successfully'}},
  549. {'worker2': {'ok': 'new rate limit set successfully'}}]
  550. * ping(destination=all, reply=False, timeout=1, limit=0)
  551. Worker returns the simple message ``"pong"``.
  552. >>> from celery.task.control import ping
  553. >>> ping(reply=True)
  554. [{'worker1': 'pong'},
  555. {'worker2': 'pong'},
  556. * revoke(destination=all, reply=False, timeout=1, limit=0)
  557. Worker simply returns ``True``.
  558. >>> from celery.task.control import revoke
  559. >>> revoke("419e46eb-cf6a-4271-86a8-442b7124132c", reply=True)
  560. [{'worker1': True},
  561. {'worker2'; True}]
  562. * You can now add your own remote control commands!
  563. Remote control commands are functions registered in the command
  564. registry. Registering a command is done using
  565. :meth:`celery.worker.control.Panel.register`:
  566. .. code-block:: python
  567. from celery.task.control import Panel
  568. @Panel.register
  569. def reset_broker_connection(panel, **kwargs):
  570. panel.listener.reset_connection()
  571. return {"ok": "connection re-established"}
  572. With this module imported in the worker, you can launch the command
  573. using ``celery.task.control.broadcast``::
  574. >>> from celery.task.control import broadcast
  575. >>> broadcast("reset_broker_connection", reply=True)
  576. [{'worker1': {'ok': 'connection re-established'},
  577. {'worker2': {'ok': 'connection re-established'}}]
  578. **TIP** You can choose the worker(s) to receive the command
  579. by using the ``destination`` argument::
  580. >>> broadcast("reset_broker_connection", destination=["worker1"])
  581. [{'worker1': {'ok': 'connection re-established'}]
  582. * New remote control command: ``dump_reserved``
  583. Dumps tasks reserved by the worker, waiting to be executed::
  584. >>> from celery.task.control import broadcast
  585. >>> broadcast("dump_reserved", reply=True)
  586. [{'myworker1': [<TaskRequest ....>]}]
  587. * New remote control command: ``dump_schedule``
  588. Dumps the workers currently registered ETA schedule.
  589. These are tasks with an ``eta`` (or ``countdown``) argument
  590. waiting to be executed by the worker.
  591. >>> from celery.task.control import broadcast
  592. >>> broadcast("dump_schedule", reply=True)
  593. [{'w1': []},
  594. {'w3': []},
  595. {'w2': ['0. 2010-05-12 11:06:00 pri0 <TaskRequest
  596. {name:"opalfeeds.tasks.refresh_feed_slice",
  597. id:"95b45760-4e73-4ce8-8eac-f100aa80273a",
  598. args:"(<Feeds freq_max:3600 freq_min:60
  599. start:2184.0 stop:3276.0>,)",
  600. kwargs:"{'page': 2}"}>']},
  601. {'w4': ['0. 2010-05-12 11:00:00 pri0 <TaskRequest
  602. {name:"opalfeeds.tasks.refresh_feed_slice",
  603. id:"c053480b-58fb-422f-ae68-8d30a464edfe",
  604. args:"(<Feeds freq_max:3600 freq_min:60
  605. start:1092.0 stop:2184.0>,)",
  606. kwargs:"{\'page\': 1}"}>',
  607. '1. 2010-05-12 11:12:00 pri0 <TaskRequest
  608. {name:"opalfeeds.tasks.refresh_feed_slice",
  609. id:"ab8bc59e-6cf8-44b8-88d0-f1af57789758",
  610. args:"(<Feeds freq_max:3600 freq_min:60
  611. start:3276.0 stop:4365>,)",
  612. kwargs:"{\'page\': 3}"}>']}]
  613. Fixes
  614. -----
  615. * Mediator thread no longer blocks for more than 1 second.
  616. With rate limits enabled and when there was a lot of remaining time,
  617. the mediator thread could block shutdown (and potentially block other
  618. jobs from coming in).
  619. * Remote rate limits was not properly applied
  620. (http://github.com/ask/celery/issues/issue/98)
  621. * Now handles exceptions with unicode messages correctly in
  622. ``TaskRequest.on_failure``.
  623. * Database backend: ``TaskMeta.result``: default value should be ``None``
  624. not empty string.
  625. 1.0.2
  626. =====
  627. :release-date: 2010-03-31 12:50 P.M CET
  628. * Deprecated: ``CELERY_BACKEND``, please use ``CELERY_RESULT_BACKEND``
  629. instead.
  630. * We now use a custom logger in tasks. This logger supports task magic
  631. keyword arguments in formats.
  632. The default format for tasks (``CELERYD_TASK_LOG_FORMAT``) now includes
  633. the id and the name of tasks so the origin of task log messages can
  634. easily be traced.
  635. Example output::
  636. [2010-03-25 13:11:20,317: INFO/PoolWorker-1]
  637. [tasks.add(a6e1c5ad-60d9-42a0-8b24-9e39363125a4)] Hello from add
  638. To revert to the previous behavior you can set::
  639. CELERYD_TASK_LOG_FORMAT = """
  640. [%(asctime)s: %(levelname)s/%(processName)s] %(message)s
  641. """.strip()
  642. * Unittests: Don't disable the django test database teardown,
  643. instead fixed the underlying issue which was caused by modifications
  644. to the ``DATABASE_NAME`` setting (http://github.com/ask/celery/issues/82).
  645. * Django Loader: New config ``CELERY_DB_REUSE_MAX`` (max number of tasks
  646. to reuse the same database connection)
  647. The default is to use a new connection for every task.
  648. We would very much like to reuse the connection, but a safe number of
  649. reuses is not known, and we don't have any way to handle the errors
  650. that might happen, which may even be database dependent.
  651. See: http://bit.ly/94fwdd
  652. * celeryd: The worker components are now configurable: ``CELERYD_POOL``,
  653. ``CELERYD_LISTENER``, ``CELERYD_MEDIATOR``, and ``CELERYD_ETA_SCHEDULER``.
  654. The default configuration is as follows:
  655. .. code-block:: python
  656. CELERYD_POOL = "celery.concurrency.processes.TaskPool"
  657. CELERYD_MEDIATOR = "celery.worker.controllers.Mediator"
  658. CELERYD_ETA_SCHEDULER = "celery.worker.controllers.ScheduleController"
  659. CELERYD_LISTENER = "celery.worker.listener.CarrotListener"
  660. The ``CELERYD_POOL`` setting makes it easy to swap out the multiprocessing
  661. pool with a threaded pool, or how about a twisted/eventlet pool?
  662. Consider the competition for the first pool plug-in started!
  663. * Debian init scripts: Use ``-a`` not ``&&``
  664. (http://github.com/ask/celery/issues/82).
  665. * Debian init scripts: Now always preserves ``$CELERYD_OPTS`` from the
  666. ``/etc/default/celeryd`` and ``/etc/default/celerybeat``.
  667. * celery.beat.Scheduler: Fixed a bug where the schedule was not properly
  668. flushed to disk if the schedule had not been properly initialized.
  669. * celerybeat: Now syncs the schedule to disk when receiving the ``SIGTERM``
  670. and ``SIGINT`` signals.
  671. * Control commands: Make sure keywords arguments are not in unicode.
  672. * ETA scheduler: Was missing a logger object, so the scheduler crashed
  673. when trying to log that a task had been revoked.
  674. * management.commands.camqadm: Fixed typo ``camqpadm`` -> ``camqadm``
  675. (http://github.com/ask/celery/issues/83).
  676. * PeriodicTask.delta_resolution: Was not working for days and hours, now fixed
  677. by rounding to the nearest day/hour.
  678. * Fixed a potential infinite loop in ``BaseAsyncResult.__eq__``, although
  679. there is no evidence that it has ever been triggered.
  680. * celeryd: Now handles messages with encoding problems by acking them and
  681. emitting an error message.
  682. 1.0.1
  683. =====
  684. :release-date: 2010-02-24 07:05 P.M CET
  685. * Tasks are now acknowledged early instead of late.
  686. This is done because messages can only be acked within the same
  687. connection channel, so if the connection is lost we would have to refetch
  688. the message again to acknowledge it.
  689. This might or might not affect you, but mostly those running tasks with a
  690. really long execution time are affected, as all tasks that has made it
  691. all the way into the pool needs to be executed before the worker can
  692. safely terminate (this is at most the number of pool workers, multiplied
  693. by the ``CELERYD_PREFETCH_MULTIPLIER`` setting.)
  694. We multiply the prefetch count by default to increase the performance at
  695. times with bursts of tasks with a short execution time. If this doesn't
  696. apply to your use case, you should be able to set the prefetch multiplier
  697. to zero, without sacrificing performance.
  698. Please note that a patch to :mod:`multiprocessing` is currently being
  699. worked on, this patch would enable us to use a better solution, and is
  700. scheduled for inclusion in the ``2.0.0`` release.
  701. * celeryd now shutdowns cleanly when receving the ``TERM`` signal.
  702. * celeryd now does a cold shutdown if the ``INT`` signal is received (Ctrl+C),
  703. this means it tries to terminate as soon as possible.
  704. * Caching of results now moved to the base backend classes, so no need
  705. to implement this functionality in the base classes.
  706. * Caches are now also limited in size, so their memory usage doesn't grow
  707. out of control.
  708. You can set the maximum number of results the cache
  709. can hold using the ``CELERY_MAX_CACHED_RESULTS`` setting (the default
  710. is five thousand results). In addition, you can refetch already retrieved
  711. results using ``backend.reload_task_result`` +
  712. ``backend.reload_taskset_result`` (that's for those who want to send
  713. results incrementally).
  714. * ``celeryd`` now works on Windows again.
  715. Note that if running with Django,
  716. you can't use ``project.settings`` as the settings module name, but the
  717. following should work::
  718. $ python manage.py celeryd --settings=settings
  719. * Execution: ``.messaging.TaskPublisher.send_task`` now
  720. incorporates all the functionality apply_async previously did.
  721. Like converting countdowns to eta, so :func:`celery.execute.apply_async` is
  722. now simply a convenient front-end to
  723. :meth:`celery.messaging.TaskPublisher.send_task`, using
  724. the task classes default options.
  725. Also :func:`celery.execute.send_task` has been
  726. introduced, which can apply tasks using just the task name (useful
  727. if the client does not have the destination task in its task registry).
  728. Example:
  729. >>> from celery.execute import send_task
  730. >>> result = send_task("celery.ping", args=[], kwargs={})
  731. >>> result.get()
  732. 'pong'
  733. * ``camqadm``: This is a new utility for command line access to the AMQP API.
  734. Excellent for deleting queues/bindings/exchanges, experimentation and
  735. testing::
  736. $ camqadm
  737. 1> help
  738. Gives an interactive shell, type ``help`` for a list of commands.
  739. When using Django, use the management command instead::
  740. $ python manage.py camqadm
  741. 1> help
  742. * Redis result backend: To conform to recent Redis API changes, the following
  743. settings has been deprecated:
  744. * ``REDIS_TIMEOUT``
  745. * ``REDIS_CONNECT_RETRY``
  746. These will emit a ``DeprecationWarning`` if used.
  747. A ``REDIS_PASSWORD`` setting has been added, so you can use the new
  748. simple authentication mechanism in Redis.
  749. * The redis result backend no longer calls ``SAVE`` when disconnecting,
  750. as this is apparently better handled by Redis itself.
  751. * If ``settings.DEBUG`` is on, celeryd now warns about the possible
  752. memory leak it can result in.
  753. * The ETA scheduler now sleeps at most two seconds between iterations.
  754. * The ETA scheduler now deletes any revoked tasks it might encounter.
  755. As revokes are not yet persistent, this is done to make sure the task
  756. is revoked even though it's currently being hold because its eta is e.g.
  757. a week into the future.
  758. * The ``task_id`` argument is now respected even if the task is executed
  759. eagerly (either using apply, or ``CELERY_ALWAYS_EAGER``).
  760. * The internal queues are now cleared if the connection is reset.
  761. * New magic keyword argument: ``delivery_info``.
  762. Used by retry() to resend the task to its original destination using the same
  763. exchange/routing_key.
  764. * Events: Fields was not passed by ``.send()`` (fixes the uuid keyerrors
  765. in celerymon)
  766. * Added ``--schedule``/``-s`` option to celeryd, so it is possible to
  767. specify a custom schedule filename when using an embedded celerybeat
  768. server (the ``-B``/``--beat``) option.
  769. * Better Python 2.4 compatibility. The test suite now passes.
  770. * task decorators: Now preserve docstring as ``cls.__doc__``, (was previously
  771. copied to ``cls.run.__doc__``)
  772. * The ``testproj`` directory has been renamed to ``tests`` and we're now using
  773. ``nose`` + ``django-nose`` for test discovery, and ``unittest2`` for test
  774. cases.
  775. * New pip requirements files available in ``contrib/requirements``.
  776. * TaskPublisher: Declarations are now done once (per process).
  777. * Added ``Task.delivery_mode`` and the ``CELERY_DEFAULT_DELIVERY_MODE``
  778. setting.
  779. These can be used to mark messages non-persistent (i.e. so they are
  780. lost if the broker is restarted).
  781. * Now have our own ``ImproperlyConfigured`` exception, instead of using the
  782. Django one.
  783. * Improvements to the debian init scripts: Shows an error if the program is
  784. not executeable. Does not modify ``CELERYD`` when using django with
  785. virtualenv.
  786. 1.0.0
  787. =====
  788. :release-date: 2010-02-10 04:00 P.M CET
  789. Backward incompatible changes
  790. -----------------------------
  791. * Celery does not support detaching anymore, so you have to use the tools
  792. available on your platform, or something like supervisord to make
  793. celeryd/celerybeat/celerymon into background processes.
  794. We've had too many problems with celeryd daemonizing itself, so it was
  795. decided it has to be removed. Example startup scripts has been added to
  796. ``contrib/``:
  797. * Debian, Ubuntu, (start-stop-daemon)
  798. ``contrib/debian/init.d/celeryd``
  799. ``contrib/debian/init.d/celerybeat``
  800. * Mac OS X launchd
  801. ``contrib/mac/org.celeryq.celeryd.plist``
  802. ``contrib/mac/org.celeryq.celerybeat.plist``
  803. ``contrib/mac/org.celeryq.celerymon.plist``
  804. * Supervisord (http://supervisord.org)
  805. ``contrib/supervisord/supervisord.conf``
  806. In addition to ``--detach``, the following program arguments has been
  807. removed: ``--uid``, ``--gid``, ``--workdir``, ``--chroot``, ``--pidfile``,
  808. ``--umask``. All good daemonization tools should support equivalent
  809. functionality, so don't worry.
  810. Also the following configuration keys has been removed:
  811. ``CELERYD_PID_FILE``, ``CELERYBEAT_PID_FILE``, ``CELERYMON_PID_FILE``.
  812. * Default celeryd loglevel is now ``WARN``, to enable the previous log level
  813. start celeryd with ``--loglevel=INFO``.
  814. * Tasks are automatically registered.
  815. This means you no longer have to register your tasks manually.
  816. You don't have to change your old code right away, as it doesn't matter if
  817. a task is registered twice.
  818. If you don't want your task to be automatically registered you can set
  819. the ``abstract`` attribute
  820. .. code-block:: python
  821. class MyTask(Task):
  822. abstract = True
  823. By using ``abstract`` only tasks subclassing this task will be automatically
  824. registered (this works like the Django ORM).
  825. If you don't want subclasses to be registered either, you can set the
  826. ``autoregister`` attribute to ``False``.
  827. Incidentally, this change also fixes the problems with automatic name
  828. assignment and relative imports. So you also don't have to specify a task name
  829. anymore if you use relative imports.
  830. * You can no longer use regular functions as tasks.
  831. This change was added
  832. because it makes the internals a lot more clean and simple. However, you can
  833. now turn functions into tasks by using the ``@task`` decorator:
  834. .. code-block:: python
  835. from celery.decorators import task
  836. @task
  837. def add(x, y):
  838. return x + y
  839. See the User Guide: :doc:`userguide/tasks` for more information.
  840. * The periodic task system has been rewritten to a centralized solution.
  841. This means ``celeryd`` no longer schedules periodic tasks by default,
  842. but a new daemon has been introduced: ``celerybeat``.
  843. To launch the periodic task scheduler you have to run celerybeat::
  844. $ celerybeat
  845. Make sure this is running on one server only, if you run it twice, all
  846. periodic tasks will also be executed twice.
  847. If you only have one worker server you can embed it into celeryd like this::
  848. $ celeryd --beat # Embed celerybeat in celeryd.
  849. * The supervisor has been removed.
  850. This means the ``-S`` and ``--supervised`` options to ``celeryd`` is
  851. no longer supported. Please use something like http://supervisord.org
  852. instead.
  853. * ``TaskSet.join`` has been removed, use ``TaskSetResult.join`` instead.
  854. * The task status ``"DONE"`` has been renamed to `"SUCCESS"`.
  855. * ``AsyncResult.is_done`` has been removed, use ``AsyncResult.successful``
  856. instead.
  857. * The worker no longer stores errors if ``Task.ignore_result`` is set, to
  858. revert to the previous behaviour set
  859. ``CELERY_STORE_ERRORS_EVEN_IF_IGNORED`` to ``True``.
  860. * The staticstics functionality has been removed in favor of events,
  861. so the ``-S`` and ``--statistics`` switches has been removed.
  862. * The module ``celery.task.strategy`` has been removed.
  863. * ``celery.discovery`` has been removed, and it's ``autodiscover`` function is
  864. now in ``celery.loaders.djangoapp``. Reason: Internal API.
  865. * ``CELERY_LOADER`` now needs loader class name in addition to module name,
  866. E.g. where you previously had: ``"celery.loaders.default"``, you now need
  867. ``"celery.loaders.default.Loader"``, using the previous syntax will result
  868. in a DeprecationWarning.
  869. * Detecting the loader is now lazy, and so is not done when importing
  870. ``celery.loaders``.
  871. To make this happen ``celery.loaders.settings`` has
  872. been renamed to ``load_settings`` and is now a function returning the
  873. settings object. ``celery.loaders.current_loader`` is now also
  874. a function, returning the current loader.
  875. So::
  876. loader = current_loader
  877. needs to be changed to::
  878. loader = current_loader()
  879. Deprecations
  880. ------------
  881. * The following configuration variables has been renamed and will be
  882. deprecated in v2.0:
  883. * CELERYD_DAEMON_LOG_FORMAT -> CELERYD_LOG_FORMAT
  884. * CELERYD_DAEMON_LOG_LEVEL -> CELERYD_LOG_LEVEL
  885. * CELERY_AMQP_CONNECTION_TIMEOUT -> CELERY_BROKER_CONNECTION_TIMEOUT
  886. * CELERY_AMQP_CONNECTION_RETRY -> CELERY_BROKER_CONNECTION_RETRY
  887. * CELERY_AMQP_CONNECTION_MAX_RETRIES -> CELERY_BROKER_CONNECTION_MAX_RETRIES
  888. * SEND_CELERY_TASK_ERROR_EMAILS -> CELERY_SEND_TASK_ERROR_EMAILS
  889. * The public api names in celery.conf has also changed to a consistent naming
  890. scheme.
  891. * We now support consuming from an arbitrary number of queues.
  892. To do this we had to rename the configuration syntax. If you use any of
  893. the custom AMQP routing options (queue/exchange/routing_key, etc), you
  894. should read the new FAQ entry: http://bit.ly/aiWoH.
  895. The previous syntax is deprecated and scheduled for removal in v2.0.
  896. * ``TaskSet.run`` has been renamed to ``TaskSet.apply_async``.
  897. ``TaskSet.run`` has now been deprecated, and is scheduled for
  898. removal in v2.0.
  899. News
  900. ----
  901. * Rate limiting support (per task type, or globally).
  902. * New periodic task system.
  903. * Automatic registration.
  904. * New cool task decorator syntax.
  905. * celeryd now sends events if enabled with the ``-E`` argument.
  906. Excellent for monitoring tools, one is already in the making
  907. (http://github.com/ask/celerymon).
  908. Current events include: worker-heartbeat,
  909. task-[received/succeeded/failed/retried],
  910. worker-online, worker-offline.
  911. * You can now delete (revoke) tasks that has already been applied.
  912. * You can now set the hostname celeryd identifies as using the ``--hostname``
  913. argument.
  914. * Cache backend now respects ``CELERY_TASK_RESULT_EXPIRES``.
  915. * Message format has been standardized and now uses ISO-8601 format
  916. for dates instead of datetime.
  917. * ``celeryd`` now responds to the ``HUP`` signal by restarting itself.
  918. * Periodic tasks are now scheduled on the clock.
  919. I.e. ``timedelta(hours=1)`` means every hour at :00 minutes, not every
  920. hour from the server starts. To revert to the previous behaviour you
  921. can set ``PeriodicTask.relative = True``.
  922. * Now supports passing execute options to a TaskSets list of args, e.g.:
  923. >>> ts = TaskSet(add, [([2, 2], {}, {"countdown": 1}),
  924. ... ([4, 4], {}, {"countdown": 2}),
  925. ... ([8, 8], {}, {"countdown": 3})])
  926. >>> ts.run()
  927. * Got a 3x performance gain by setting the prefetch count to four times the
  928. concurrency, (from an average task round-trip of 0.1s to 0.03s!).
  929. A new setting has been added: ``CELERYD_PREFETCH_MULTIPLIER``, which
  930. is set to ``4`` by default.
  931. * Improved support for webhook tasks.
  932. ``celery.task.rest`` is now deprecated, replaced with the new and shiny
  933. :mod:`celery.task.http`. With more reflective names, sensible interface,
  934. and it's possible to override the methods used to perform HTTP requests.
  935. * The results of tasksets are now cached by storing it in the result
  936. backend.
  937. Changes
  938. -------
  939. * Now depends on carrot >= 0.8.1
  940. * New dependencies: billiard, python-dateutil, django-picklefield
  941. * No longer depends on python-daemon
  942. * The ``uuid`` distribution is added as a dependency when running Python 2.4.
  943. * Now remembers the previously detected loader by keeping it in
  944. the ``CELERY_LOADER`` environment variable.
  945. This may help on windows where fork emulation is used.
  946. * ETA no longer sends datetime objects, but uses ISO 8601 date format in a
  947. string for better compatibility with other platforms.
  948. * No longer sends error mails for retried tasks.
  949. * Task can now override the backend used to store results.
  950. * Refactored the ExecuteWrapper, ``apply`` and ``CELERY_ALWAYS_EAGER`` now
  951. also executes the task callbacks and signals.
  952. * Now using a proper scheduler for the tasks with an ETA.
  953. This means waiting eta tasks are sorted by time, so we don't have
  954. to poll the whole list all the time.
  955. * Now also imports modules listed in CELERY_IMPORTS when running
  956. with django (as documented).
  957. * Loglevel for stdout/stderr changed from INFO to ERROR
  958. * ImportErrors are now properly propogated when autodiscovering tasks.
  959. * You can now use ``celery.messaging.establish_connection`` to establish a
  960. connection to the broker.
  961. * When running as a separate service the periodic task scheduler does some
  962. smart moves to not poll too regularly.
  963. If you need faster poll times you can lower the value
  964. of ``CELERYBEAT_MAX_LOOP_INTERVAL``.
  965. * You can now change periodic task intervals at runtime, by making
  966. ``run_every`` a property, or subclassing ``PeriodicTask.is_due``.
  967. * The worker now supports control commands enabled through the use of a
  968. broadcast queue, you can remotely revoke tasks or set the rate limit for
  969. a task type. See :mod:`celery.task.control`.
  970. * The services now sets informative process names (as shown in ``ps``
  971. listings) if the :mod:`setproctitle` module is installed.
  972. * :exc:`celery.exceptions.NotRegistered` now inherits from :exc:`KeyError`,
  973. and ``TaskRegistry.__getitem__``+``pop`` raises ``NotRegistered`` instead
  974. * You can set the loader via the ``CELERY_LOADER`` environment variable.
  975. * You can now set ``CELERY_IGNORE_RESULT`` to ignore task results by default
  976. (if enabled, tasks doesn't save results or errors to the backend used).
  977. * celeryd now correctly handles malformed messages by throwing away and
  978. acknowledging the message, instead of crashing.
  979. Bugs
  980. ----
  981. * Fixed a race condition that could happen while storing task results in the
  982. database.
  983. Documentation
  984. -------------
  985. * Reference now split into two sections; API reference and internal module
  986. reference.
  987. 0.8.4
  988. =====
  989. :release-date: 2010-02-05 01:52 P.M CEST
  990. * Now emits a warning if the --detach argument is used.
  991. --detach should not be used anymore, as it has several not easily fixed
  992. bugs related to it. Instead, use something like start-stop-daemon,
  993. supervisord or launchd (os x).
  994. * Make sure logger class is process aware, even if running Python >= 2.6.
  995. * Error e-mails are not sent anymore when the task is retried.
  996. 0.8.3
  997. =====
  998. :release-date: 2009-12-22 09:43 A.M CEST
  999. * Fixed a possible race condition that could happen when storing/querying
  1000. task results using the the database backend.
  1001. * Now has console script entry points in the setup.py file, so tools like
  1002. buildout will correctly install the programs celerybin and celeryinit.
  1003. 0.8.2
  1004. =====
  1005. :release-date: 2009-11-20 03:40 P.M CEST
  1006. * QOS Prefetch count was not applied properly, as it was set for every message
  1007. received (which apparently behaves like, "receive one more"), instead of only
  1008. set when our wanted value cahnged.
  1009. 0.8.1
  1010. =================================
  1011. :release-date: 2009-11-16 05:21 P.M CEST
  1012. Very important note
  1013. -------------------
  1014. This release (with carrot 0.8.0) enables AMQP QoS (quality of service), which
  1015. means the workers will only receive as many messages as it can handle at a
  1016. time. As with any release, you should test this version upgrade on your
  1017. development servers before rolling it out to production!
  1018. Important changes
  1019. -----------------
  1020. * If you're using Python < 2.6 and you use the multiprocessing backport, then
  1021. multiprocessing version 2.6.2.1 is required.
  1022. * All AMQP_* settings has been renamed to BROKER_*, and in addition
  1023. AMQP_SERVER has been renamed to BROKER_HOST, so before where you had::
  1024. AMQP_SERVER = "localhost"
  1025. AMQP_PORT = 5678
  1026. AMQP_USER = "myuser"
  1027. AMQP_PASSWORD = "mypassword"
  1028. AMQP_VHOST = "celery"
  1029. You need to change that to::
  1030. BROKER_HOST = "localhost"
  1031. BROKER_PORT = 5678
  1032. BROKER_USER = "myuser"
  1033. BROKER_PASSWORD = "mypassword"
  1034. BROKER_VHOST = "celery"
  1035. * Custom carrot backends now need to include the backend class name, so before
  1036. where you had::
  1037. CARROT_BACKEND = "mycustom.backend.module"
  1038. you need to change it to::
  1039. CARROT_BACKEND = "mycustom.backend.module.Backend"
  1040. where ``Backend`` is the class name. This is probably ``"Backend"``, as
  1041. that was the previously implied name.
  1042. * New version requirement for carrot: 0.8.0
  1043. Changes
  1044. -------
  1045. * Incorporated the multiprocessing backport patch that fixes the
  1046. ``processName`` error.
  1047. * Ignore the result of PeriodicTask's by default.
  1048. * Added a Redis result store backend
  1049. * Allow /etc/default/celeryd to define additional options for the celeryd init
  1050. script.
  1051. * MongoDB periodic tasks issue when using different time than UTC fixed.
  1052. * Windows specific: Negate test for available os.fork (thanks miracle2k)
  1053. * Now tried to handle broken PID files.
  1054. * Added a Django test runner to contrib that sets CELERY_ALWAYS_EAGER = True for testing with the database backend
  1055. * Added a CELERY_CACHE_BACKEND setting for using something other than the django-global cache backend.
  1056. * Use custom implementation of functools.partial (curry) for Python 2.4 support
  1057. (Probably still problems with running on 2.4, but it will eventually be
  1058. supported)
  1059. * Prepare exception to pickle when saving RETRY status for all backends.
  1060. * SQLite no concurrency limit should only be effective if the db backend is used.
  1061. 0.8.0
  1062. =====
  1063. :release-date: 2009-09-22 03:06 P.M CEST
  1064. Backward incompatible changes
  1065. -----------------------------
  1066. * Add traceback to result value on failure.
  1067. **NOTE** If you use the database backend you have to re-create the
  1068. database table ``celery_taskmeta``.
  1069. Contact the mailinglist or IRC channel listed in README for help
  1070. doing this.
  1071. * Database tables are now only created if the database backend is used,
  1072. so if you change back to the database backend at some point,
  1073. be sure to initialize tables (django: ``syncdb``, python: ``celeryinit``).
  1074. (Note: This is only the case when using Django 1.1 or higher)
  1075. * Now depends on ``carrot`` version 0.6.0.
  1076. * Now depends on python-daemon 1.4.8
  1077. Important changes
  1078. -----------------
  1079. * Celery can now be used in pure Python (outside of a Django project).
  1080. This means celery is no longer Django specific.
  1081. For more information see the FAQ entry
  1082. `Can I use celery without Django?`_.
  1083. .. _`Can I use celery without Django?`:
  1084. http://ask.github.com/celery/faq.html#can-i-use-celery-without-django
  1085. * Celery now supports task retries.
  1086. See `Cookbook: Retrying Tasks`_ for more information.
  1087. .. _`Cookbook: Retrying Tasks`:
  1088. http://ask.github.com/celery/cookbook/task-retries.html
  1089. * We now have an AMQP result store backend.
  1090. It uses messages to publish task return value and status. And it's
  1091. incredibly fast!
  1092. See http://github.com/ask/celery/issues/closed#issue/6 for more info!
  1093. * AMQP QoS (prefetch count) implemented:
  1094. This to not receive more messages than we can handle.
  1095. * Now redirects stdout/stderr to the celeryd logfile when detached
  1096. * Now uses ``inspect.getargspec`` to only pass default arguments
  1097. the task supports.
  1098. * Add Task.on_success, .on_retry, .on_failure handlers
  1099. See :meth:`celery.task.base.Task.on_success`,
  1100. :meth:`celery.task.base.Task.on_retry`,
  1101. :meth:`celery.task.base.Task.on_failure`,
  1102. * ``celery.utils.gen_unique_id``: Workaround for
  1103. http://bugs.python.org/issue4607
  1104. * You can now customize what happens at worker start, at process init, etc
  1105. by creating your own loaders. (see :mod:`celery.loaders.default`,
  1106. :mod:`celery.loaders.djangoapp`, :mod:`celery.loaders`.)
  1107. * Support for multiple AMQP exchanges and queues.
  1108. This feature misses documentation and tests, so anyone interested
  1109. is encouraged to improve this situation.
  1110. * celeryd now survives a restart of the AMQP server!
  1111. Automatically re-establish AMQP broker connection if it's lost.
  1112. New settings:
  1113. * AMQP_CONNECTION_RETRY
  1114. Set to ``True`` to enable connection retries.
  1115. * AMQP_CONNECTION_MAX_RETRIES.
  1116. Maximum number of restarts before we give up. Default: ``100``.
  1117. News
  1118. ----
  1119. * Fix an incompatibility between python-daemon and multiprocessing,
  1120. which resulted in the ``[Errno 10] No child processes`` problem when
  1121. detaching.
  1122. * Fixed a possible DjangoUnicodeDecodeError being raised when saving pickled
  1123. data to Django's memcached cache backend.
  1124. * Better Windows compatibility.
  1125. * New version of the pickled field (taken from
  1126. http://www.djangosnippets.org/snippets/513/)
  1127. * New signals introduced: ``task_sent``, ``task_prerun`` and
  1128. ``task_postrun``, see :mod:`celery.signals` for more information.
  1129. * ``TaskSetResult.join`` caused ``TypeError`` when ``timeout=None``.
  1130. Thanks Jerzy Kozera. Closes #31
  1131. * ``views.apply`` should return ``HttpResponse`` instance.
  1132. Thanks to Jerzy Kozera. Closes #32
  1133. * ``PeriodicTask``: Save conversion of ``run_every`` from ``int``
  1134. to ``timedelta`` to the class attribute instead of on the instance.
  1135. * Exceptions has been moved to ``celery.exceptions``, but are still
  1136. available in the previous module.
  1137. * Try to rollback transaction and retry saving result if an error happens
  1138. while setting task status with the database backend.
  1139. * jail() refactored into :class:`celery.execute.ExecuteWrapper`.
  1140. * ``views.apply`` now correctly sets mimetype to "application/json"
  1141. * ``views.task_status`` now returns exception if status is RETRY
  1142. * ``views.task_status`` now returns traceback if status is "FAILURE"
  1143. or "RETRY"
  1144. * Documented default task arguments.
  1145. * Add a sensible __repr__ to ExceptionInfo for easier debugging
  1146. * Fix documentation typo ``.. import map`` -> ``.. import dmap``.
  1147. Thanks mikedizon
  1148. 0.6.0
  1149. =====
  1150. :release-date: 2009-08-07 06:54 A.M CET
  1151. Important changes
  1152. -----------------
  1153. * Fixed a bug where tasks raising unpickleable exceptions crashed pool
  1154. workers. So if you've had pool workers mysteriously dissapearing, or
  1155. problems with celeryd stopping working, this has been fixed in this
  1156. version.
  1157. * Fixed a race condition with periodic tasks.
  1158. * The task pool is now supervised, so if a pool worker crashes,
  1159. goes away or stops responding, it is automatically replaced with
  1160. a new one.
  1161. * Task.name is now automatically generated out of class module+name, e.g.
  1162. ``"djangotwitter.tasks.UpdateStatusesTask"``. Very convenient. No idea why
  1163. we didn't do this before. Some documentation is updated to not manually
  1164. specify a task name.
  1165. News
  1166. ----
  1167. * Tested with Django 1.1
  1168. * New Tutorial: Creating a click counter using carrot and celery
  1169. * Database entries for periodic tasks are now created at ``celeryd``
  1170. startup instead of for each check (which has been a forgotten TODO/XXX
  1171. in the code for a long time)
  1172. * New settings variable: ``CELERY_TASK_RESULT_EXPIRES``
  1173. Time (in seconds, or a `datetime.timedelta` object) for when after
  1174. stored task results are deleted. For the moment this only works for the
  1175. database backend.
  1176. * ``celeryd`` now emits a debug log message for which periodic tasks
  1177. has been launched.
  1178. * The periodic task table is now locked for reading while getting
  1179. periodic task status. (MySQL only so far, seeking patches for other
  1180. engines)
  1181. * A lot more debugging information is now available by turning on the
  1182. ``DEBUG`` loglevel (``--loglevel=DEBUG``).
  1183. * Functions/methods with a timeout argument now works correctly.
  1184. * New: ``celery.strategy.even_time_distribution``:
  1185. With an iterator yielding task args, kwargs tuples, evenly distribute
  1186. the processing of its tasks throughout the time window available.
  1187. * Log message ``Unknown task ignored...`` now has loglevel ``ERROR``
  1188. * Log message ``"Got task from broker"`` is now emitted for all tasks, even if
  1189. the task has an ETA (estimated time of arrival). Also the message now
  1190. includes the ETA for the task (if any).
  1191. * Acknowledgement now happens in the pool callback. Can't do ack in the job
  1192. target, as it's not pickleable (can't share AMQP connection, etc)).
  1193. * Added note about .delay hanging in README
  1194. * Tests now passing in Django 1.1
  1195. * Fixed discovery to make sure app is in INSTALLED_APPS
  1196. * Previously overrided pool behaviour (process reap, wait until pool worker
  1197. available, etc.) is now handled by ``multiprocessing.Pool`` itself.
  1198. * Convert statistics data to unicode for use as kwargs. Thanks Lucy!
  1199. 0.4.1
  1200. =====
  1201. :release-date: 2009-07-02 01:42 P.M CET
  1202. * Fixed a bug with parsing the message options (``mandatory``,
  1203. ``routing_key``, ``priority``, ``immediate``)
  1204. 0.4.0
  1205. =====
  1206. :release-date: 2009-07-01 07:29 P.M CET
  1207. * Adds eager execution. ``celery.execute.apply``|``Task.apply`` executes the
  1208. function blocking until the task is done, for API compatiblity it
  1209. returns an ``celery.result.EagerResult`` instance. You can configure
  1210. celery to always run tasks locally by setting the
  1211. ``CELERY_ALWAYS_EAGER`` setting to ``True``.
  1212. * Now depends on ``anyjson``.
  1213. * 99% coverage using python ``coverage`` 3.0.
  1214. 0.3.20
  1215. ======
  1216. :release-date: 2009-06-25 08:42 P.M CET
  1217. * New arguments to ``apply_async`` (the advanced version of
  1218. ``delay_task``), ``countdown`` and ``eta``;
  1219. >>> # Run 10 seconds into the future.
  1220. >>> res = apply_async(MyTask, countdown=10);
  1221. >>> # Run 1 day from now
  1222. >>> res = apply_async(MyTask, eta=datetime.now() +
  1223. ... timedelta(days=1)
  1224. * Now unlinks the pidfile if it's stale.
  1225. * Lots of more tests.
  1226. * Now compatible with carrot >= 0.5.0.
  1227. * **IMPORTANT** The ``subtask_ids`` attribute on the ``TaskSetResult``
  1228. instance has been removed. To get this information instead use:
  1229. >>> subtask_ids = [subtask.task_id for subtask in ts_res.subtasks]
  1230. * ``Taskset.run()`` now respects extra message options from the task class.
  1231. * Task: Add attribute ``ignore_result``: Don't store the status and
  1232. return value. This means you can't use the
  1233. ``celery.result.AsyncResult`` to check if the task is
  1234. done, or get its return value. Only use if you need the performance
  1235. and is able live without these features. Any exceptions raised will
  1236. store the return value/status as usual.
  1237. * Task: Add attribute ``disable_error_emails`` to disable sending error
  1238. emails for that task.
  1239. * Should now work on Windows (although running in the background won't
  1240. work, so using the ``--detach`` argument results in an exception
  1241. being raised.)
  1242. * Added support for statistics for profiling and monitoring.
  1243. To start sending statistics start ``celeryd`` with the
  1244. ``--statistics`` option. Then after a while you can dump the results
  1245. by running ``python manage.py celerystats``. See
  1246. ``celery.monitoring`` for more information.
  1247. * The celery daemon can now be supervised (i.e it is automatically
  1248. restarted if it crashes). To use this start celeryd with the
  1249. ``--supervised`` option (or alternatively ``-S``).
  1250. * views.apply: View applying a task. Example::
  1251. http://e.com/celery/apply/task_name/arg1/arg2//?kwarg1=a&kwarg2=b
  1252. **NOTE** Use with caution, preferably not make this publicly
  1253. accessible without ensuring your code is safe!
  1254. * Refactored ``celery.task``. It's now split into three modules:
  1255. * celery.task
  1256. Contains ``apply_async``, ``delay_task``, ``discard_all``, and task
  1257. shortcuts, plus imports objects from ``celery.task.base`` and
  1258. ``celery.task.builtins``
  1259. * celery.task.base
  1260. Contains task base classes: ``Task``, ``PeriodicTask``,
  1261. ``TaskSet``, ``AsynchronousMapTask``, ``ExecuteRemoteTask``.
  1262. * celery.task.builtins
  1263. Built-in tasks: ``PingTask``, ``DeleteExpiredTaskMetaTask``.
  1264. 0.3.7
  1265. =====
  1266. :release-date: 2008-06-16 11:41 P.M CET
  1267. * **IMPORTANT** Now uses AMQP's ``basic.consume`` instead of
  1268. ``basic.get``. This means we're no longer polling the broker for
  1269. new messages.
  1270. * **IMPORTANT** Default concurrency limit is now set to the number of CPUs
  1271. available on the system.
  1272. * **IMPORTANT** ``tasks.register``: Renamed ``task_name`` argument to
  1273. ``name``, so
  1274. >>> tasks.register(func, task_name="mytask")
  1275. has to be replaced with:
  1276. >>> tasks.register(func, name="mytask")
  1277. * The daemon now correctly runs if the pidlock is stale.
  1278. * Now compatible with carrot 0.4.5
  1279. * Default AMQP connnection timeout is now 4 seconds.
  1280. * ``AsyncResult.read()`` was always returning ``True``.
  1281. * Only use README as long_description if the file exists so easy_install
  1282. doesn't break.
  1283. * ``celery.view``: JSON responses now properly set its mime-type.
  1284. * ``apply_async`` now has a ``connection`` keyword argument so you
  1285. can re-use the same AMQP connection if you want to execute
  1286. more than one task.
  1287. * Handle failures in task_status view such that it won't throw 500s.
  1288. * Fixed typo ``AMQP_SERVER`` in documentation to ``AMQP_HOST``.
  1289. * Worker exception e-mails sent to admins now works properly.
  1290. * No longer depends on ``django``, so installing ``celery`` won't affect
  1291. the preferred Django version installed.
  1292. * Now works with PostgreSQL (psycopg2) again by registering the
  1293. ``PickledObject`` field.
  1294. * ``celeryd``: Added ``--detach`` option as an alias to ``--daemon``, and
  1295. it's the term used in the documentation from now on.
  1296. * Make sure the pool and periodic task worker thread is terminated
  1297. properly at exit. (So ``Ctrl-C`` works again).
  1298. * Now depends on ``python-daemon``.
  1299. * Removed dependency to ``simplejson``
  1300. * Cache Backend: Re-establishes connection for every task process
  1301. if the Django cache backend is memcached/libmemcached.
  1302. * Tyrant Backend: Now re-establishes the connection for every task
  1303. executed.
  1304. 0.3.3
  1305. =====
  1306. :release-date: 2009-06-08 01:07 P.M CET
  1307. * The ``PeriodicWorkController`` now sleeps for 1 second between checking
  1308. for periodic tasks to execute.
  1309. 0.3.2
  1310. =====
  1311. :release-date: 2009-06-08 01:07 P.M CET
  1312. * celeryd: Added option ``--discard``: Discard (delete!) all waiting
  1313. messages in the queue.
  1314. * celeryd: The ``--wakeup-after`` option was not handled as a float.
  1315. 0.3.1
  1316. =====
  1317. :release-date: 2009-06-08 01:07 P.M CET
  1318. * The `PeriodicTask`` worker is now running in its own thread instead
  1319. of blocking the ``TaskController`` loop.
  1320. * Default ``QUEUE_WAKEUP_AFTER`` has been lowered to ``0.1`` (was ``0.3``)
  1321. 0.3.0
  1322. =====
  1323. :release-date: 2009-06-08 12:41 P.M CET
  1324. **NOTE** This is a development version, for the stable release, please
  1325. see versions 0.2.x.
  1326. **VERY IMPORTANT:** Pickle is now the encoder used for serializing task
  1327. arguments, so be sure to flush your task queue before you upgrade.
  1328. * **IMPORTANT** TaskSet.run() now returns a celery.result.TaskSetResult
  1329. instance, which lets you inspect the status and return values of a
  1330. taskset as it was a single entity.
  1331. * **IMPORTANT** Celery now depends on carrot >= 0.4.1.
  1332. * The celery daemon now sends task errors to the registered admin e-mails.
  1333. To turn off this feature, set ``SEND_CELERY_TASK_ERROR_EMAILS`` to
  1334. ``False`` in your ``settings.py``. Thanks to Grégoire Cachet.
  1335. * You can now run the celery daemon by using ``manage.py``::
  1336. $ python manage.py celeryd
  1337. Thanks to Grégoire Cachet.
  1338. * Added support for message priorities, topic exchanges, custom routing
  1339. keys for tasks. This means we have introduced
  1340. ``celery.task.apply_async``, a new way of executing tasks.
  1341. You can use ``celery.task.delay`` and ``celery.Task.delay`` like usual, but
  1342. if you want greater control over the message sent, you want
  1343. ``celery.task.apply_async`` and ``celery.Task.apply_async``.
  1344. This also means the AMQP configuration has changed. Some settings has
  1345. been renamed, while others are new::
  1346. CELERY_AMQP_EXCHANGE
  1347. CELERY_AMQP_PUBLISHER_ROUTING_KEY
  1348. CELERY_AMQP_CONSUMER_ROUTING_KEY
  1349. CELERY_AMQP_CONSUMER_QUEUE
  1350. CELERY_AMQP_EXCHANGE_TYPE
  1351. See the entry `Can I send some tasks to only some servers?`_ in the
  1352. `FAQ`_ for more information.
  1353. .. _`Can I send some tasks to only some servers?`:
  1354. http://bit.ly/celery_AMQP_routing
  1355. .. _`FAQ`: http://ask.github.com/celery/faq.html
  1356. * Task errors are now logged using loglevel ``ERROR`` instead of ``INFO``,
  1357. and backtraces are dumped. Thanks to Grégoire Cachet.
  1358. * Make every new worker process re-establish it's Django DB connection,
  1359. this solving the "MySQL connection died?" exceptions.
  1360. Thanks to Vitaly Babiy and Jirka Vejrazka.
  1361. * **IMOPORTANT** Now using pickle to encode task arguments. This means you
  1362. now can pass complex python objects to tasks as arguments.
  1363. * Removed dependency to ``yadayada``.
  1364. * Added a FAQ, see ``docs/faq.rst``.
  1365. * Now converts any unicode keys in task ``kwargs`` to regular strings.
  1366. Thanks Vitaly Babiy.
  1367. * Renamed the ``TaskDaemon`` to ``WorkController``.
  1368. * ``celery.datastructures.TaskProcessQueue`` is now renamed to
  1369. ``celery.pool.TaskPool``.
  1370. * The pool algorithm has been refactored for greater performance and
  1371. stability.
  1372. 0.2.0
  1373. =====
  1374. :release-date: 2009-05-20 05:14 P.M CET
  1375. * Final release of 0.2.0
  1376. * Compatible with carrot version 0.4.0.
  1377. * Fixes some syntax errors related to fetching results
  1378. from the database backend.
  1379. 0.2.0-pre3
  1380. ==========
  1381. :release-date: 2009-05-20 05:14 P.M CET
  1382. * *Internal release*. Improved handling of unpickled exceptions,
  1383. ``get_result`` now tries to recreate something looking like the
  1384. original exception.
  1385. 0.2.0-pre2
  1386. ==========
  1387. :release-date: 2009-05-20 01:56 P.M CET
  1388. * Now handles unpickleable exceptions (like the dynimically generated
  1389. subclasses of ``django.core.exception.MultipleObjectsReturned``).
  1390. 0.2.0-pre1
  1391. ==========
  1392. :release-date: 2009-05-20 12:33 P.M CET
  1393. * It's getting quite stable, with a lot of new features, so bump
  1394. version to 0.2. This is a pre-release.
  1395. * ``celery.task.mark_as_read()`` and ``celery.task.mark_as_failure()`` has
  1396. been removed. Use ``celery.backends.default_backend.mark_as_read()``,
  1397. and ``celery.backends.default_backend.mark_as_failure()`` instead.
  1398. 0.1.15
  1399. ======
  1400. :release-date: 2009-05-19 04:13 P.M CET
  1401. * The celery daemon was leaking AMQP connections, this should be fixed,
  1402. if you have any problems with too many files open (like ``emfile``
  1403. errors in ``rabbit.log``, please contact us!
  1404. 0.1.14
  1405. ======
  1406. :release-date: 2009-05-19 01:08 P.M CET
  1407. * Fixed a syntax error in the ``TaskSet`` class. (No such variable
  1408. ``TimeOutError``).
  1409. 0.1.13
  1410. ======
  1411. :release-date: 2009-05-19 12:36 P.M CET
  1412. * Forgot to add ``yadayada`` to install requirements.
  1413. * Now deletes all expired task results, not just those marked as done.
  1414. * Able to load the Tokyo Tyrant backend class without django
  1415. configuration, can specify tyrant settings directly in the class
  1416. constructor.
  1417. * Improved API documentation
  1418. * Now using the Sphinx documentation system, you can build
  1419. the html documentation by doing ::
  1420. $ cd docs
  1421. $ make html
  1422. and the result will be in ``docs/.build/html``.
  1423. 0.1.12
  1424. ======
  1425. :release-date: 2009-05-18 04:38 P.M CET
  1426. * ``delay_task()`` etc. now returns ``celery.task.AsyncResult`` object,
  1427. which lets you check the result and any failure that might have
  1428. happened. It kind of works like the ``multiprocessing.AsyncResult``
  1429. class returned by ``multiprocessing.Pool.map_async``.
  1430. * Added dmap() and dmap_async(). This works like the
  1431. ``multiprocessing.Pool`` versions except they are tasks
  1432. distributed to the celery server. Example:
  1433. >>> from celery.task import dmap
  1434. >>> import operator
  1435. >>> dmap(operator.add, [[2, 2], [4, 4], [8, 8]])
  1436. >>> [4, 8, 16]
  1437. >>> from celery.task import dmap_async
  1438. >>> import operator
  1439. >>> result = dmap_async(operator.add, [[2, 2], [4, 4], [8, 8]])
  1440. >>> result.ready()
  1441. False
  1442. >>> time.sleep(1)
  1443. >>> result.ready()
  1444. True
  1445. >>> result.result
  1446. [4, 8, 16]
  1447. * Refactored the task metadata cache and database backends, and added
  1448. a new backend for Tokyo Tyrant. You can set the backend in your django
  1449. settings file. e.g::
  1450. CELERY_RESULT_BACKEND = "database"; # Uses the database
  1451. CELERY_RESULT_BACKEND = "cache"; # Uses the django cache framework
  1452. CELERY_RESULT_BACKEND = "tyrant"; # Uses Tokyo Tyrant
  1453. TT_HOST = "localhost"; # Hostname for the Tokyo Tyrant server.
  1454. TT_PORT = 6657; # Port of the Tokyo Tyrant server.
  1455. 0.1.11
  1456. ======
  1457. :release-date: 2009-05-12 02:08 P.M CET
  1458. * The logging system was leaking file descriptors, resulting in
  1459. servers stopping with the EMFILES (too many open files) error. (fixed)
  1460. 0.1.10
  1461. ======
  1462. :release-date: 2009-05-11 12:46 P.M CET
  1463. * Tasks now supports both positional arguments and keyword arguments.
  1464. * Requires carrot 0.3.8.
  1465. * The daemon now tries to reconnect if the connection is lost.
  1466. 0.1.8
  1467. =====
  1468. :release-date: 2009-05-07 12:27 P.M CET
  1469. * Better test coverage
  1470. * More documentation
  1471. * celeryd doesn't emit ``Queue is empty`` message if
  1472. ``settings.CELERYD_EMPTY_MSG_EMIT_EVERY`` is 0.
  1473. 0.1.7
  1474. =====
  1475. :release-date: 2009-04-30 1:50 P.M CET
  1476. * Added some unittests
  1477. * Can now use the database for task metadata (like if the task has
  1478. been executed or not). Set ``settings.CELERY_TASK_META``
  1479. * Can now run ``python setup.py test`` to run the unittests from
  1480. within the ``tests`` project.
  1481. * Can set the AMQP exchange/routing key/queue using
  1482. ``settings.CELERY_AMQP_EXCHANGE``, ``settings.CELERY_AMQP_ROUTING_KEY``,
  1483. and ``settings.CELERY_AMQP_CONSUMER_QUEUE``.
  1484. 0.1.6
  1485. =====
  1486. :release-date: 2009-04-28 2:13 P.M CET
  1487. * Introducing ``TaskSet``. A set of subtasks is executed and you can
  1488. find out how many, or if all them, are done (excellent for progress
  1489. bars and such)
  1490. * Now catches all exceptions when running ``Task.__call__``, so the
  1491. daemon doesn't die. This does't happen for pure functions yet, only
  1492. ``Task`` classes.
  1493. * ``autodiscover()`` now works with zipped eggs.
  1494. * celeryd: Now adds curernt working directory to ``sys.path`` for
  1495. convenience.
  1496. * The ``run_every`` attribute of ``PeriodicTask`` classes can now be a
  1497. ``datetime.timedelta()`` object.
  1498. * celeryd: You can now set the ``DJANGO_PROJECT_DIR`` variable
  1499. for ``celeryd`` and it will add that to ``sys.path`` for easy launching.
  1500. * Can now check if a task has been executed or not via HTTP.
  1501. * You can do this by including the celery ``urls.py`` into your project,
  1502. >>> url(r'^celery/$', include("celery.urls"))
  1503. then visiting the following url,::
  1504. http://mysite/celery/$task_id/done/
  1505. this will return a JSON dictionary like e.g:
  1506. >>> {"task": {"id": $task_id, "executed": true}}
  1507. * ``delay_task`` now returns string id, not ``uuid.UUID`` instance.
  1508. * Now has ``PeriodicTasks``, to have ``cron`` like functionality.
  1509. * Project changed name from ``crunchy`` to ``celery``. The details of
  1510. the name change request is in ``docs/name_change_request.txt``.
  1511. 0.1.0
  1512. =====
  1513. :release-date: 2009-04-24 11:28 A.M CET
  1514. * Initial release