django异步任务celery，分布式队列(redis/rabbitmq)

         异步任务用提celery.

         队列用的是redis,rabbitmq两种

        

        celery思路：

              (1)引用celery

    #coding:utf-8

from __future__ import absolute_import
from celery import Celery

app = Celery('T1',include=['T1.tasks','T1.call_tasks'])
app.config_from_object('T1.config')

if __name__ ==  '__main__':
    app.start()

              (2)引用celery配置

#coding:utf-8

from __future__ import absolute_import
from datetime import timedelta
from  time import  timezone

CELERY_RESULT_BACKEND = 'redis://127.0.0.1:6379/1'
BROKER_URL = 'redis://127.0.0.1:6379/0'

CELERY_TASK_SERIALIZER='json'
CELERY_ACCEPT_CONTENT=['json']
CELERY_RESULT_SERIALIZER='json'
CELERY_TIMEZONE='Europe/Oslo'
CELERY_ENABLE_UTC=True

CELERYBEAT_SCHEDULE = {
    'add-every-5-seconds':{
        'task':'T1.tasks.add',
        'schedule':timedelta(seconds=5),
        'args':(16,16),
    },
    'add-every-10-seconds':{
        'task':'T1.call_tasks.call_tasks',
        'schedule':timedelta(seconds=10),
        'args':(16,16),
    },
    'add-every-10-seconds':{
        'task':'T1.call_tasks.call_task2',
        'schedule':timedelta(seconds=10),
        'args':(16,16),
    }


}

              (3)

    把任务放进celery

    #coding:utf-8

from __future__ import absolute_import
from T1.celery import app
# brokers = 'redis://127.0.0.1:6379/0'
# backend = 'redis://127.0.0.1:6379/1'
#
# app = Celery('tasks',broker=brokers,backend=backend)

@app.task
def add(x,y):
    z = x+y
    return z

              (4)

   调用的任务也放进celery并定时跑

#coding:utf-8
from __future__ import absolute_import
from T1.tasks import add
from T1.celery import app
import json


@app.task
def call_tasks(n=100,m=200):
    r = add.delay(2,6)
    with open('/Users/admin/devops/Django_T1/celery.txt','wb+') as f:
        f.write(str(r.ready())+str("结果是："+str(r.get()))+"\r\n")

@app.task
def call_task2(n=100,m=200):
    r = add.delay(3,6)
    with open('/Users/admin/devops/Django_T1/celery.txt','wb+') as f:
        f.write(str(r.ready())+str("结果是："+str(r.get()))+"\r\n")

 

      启动redis

      运行celery

      celery -A T1 worker -B -l info

     

      运行结果：

      False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

False结果是：9

True结果是：9

False结果是：9

False结果是：9

False结果是：9

 

         生产者，消费者 放在redis是一次消费，一次取了，就没了

 

         以上是基于redis的队列。下面是基于rabbitmq的队列

         思路是一样的，只是安装rabbitmq比较费点时间

         rabbitmq 安装及异常处理

        http://hugoren.iteye.com/blogs/2303162

        

reference:

http://www.jianshu.com/p/1840035cb510

http://docs.jinkan.org/docs/celery/index.html