我正在使用 Flask 和 Celery，尝试锁定特定任务，以便一次只能运行一个。在 celery 文档中，它提供了一个示例来实现这个 Celery docs, Ensuring a task is only executed one at a time。虽然这个示例是针对 Django 的，但我已经尽力将其转换为适用于 Flask 的代码。然而，我仍然发现有锁的 myTask1 可以运行多次。
有一件事情不太清楚，就是我是否正确地使用了缓存。我以前从未使用过，所有内容都很新鲜。文档中提到但没有解释的一件事是：
这里还有一张截图，你可以看到我运行了
有一件事情不太清楚，就是我是否正确地使用了缓存。我以前从未使用过，所有内容都很新鲜。文档中提到但没有解释的一件事是：
文档说明：
为了使此功能正常工作，您需要使用缓存后端，其中 .add 操作是原子操作。memcached 已知可用于此目的。
我不太确定这意味着什么，我是否应该将缓存与数据库一起使用，如果是的话，我该如何做？我正在使用mongodb。在我的代码中，我只是为缓存设置了这个 cache = Cache(app, config={'CACHE_TYPE': 'simple'})，因为这是 Flask-Cache 文档中提到的Flask-Cache Docs。
另一件让我不清楚的事情是，由于我是从我的 Flask 路由 task1 中调用 myTask1，所以我是否需要做任何不同的事情。
以下是我正在使用的代码示例。
from flask import (Flask, render_template, flash, redirect,
url_for, session, logging, request, g, render_template_string, jsonify)
from flask_caching import Cache
from contextlib import contextmanager
from celery import Celery
from Flask_celery import make_celery
from celery.result import AsyncResult
from celery.utils.log import get_task_logger
from celery.five import monotonic
from flask_pymongo import PyMongo
from hashlib import md5
import pymongo
import time
app = Flask(__name__)
cache = Cache(app, config={'CACHE_TYPE': 'simple'})
app.config['SECRET_KEY']= 'super secret key for me123456789987654321'
######################
# MONGODB SETUP
#####################
app.config['MONGO_HOST'] = 'localhost'
app.config['MONGO_DBNAME'] = 'celery-test-db'
app.config["MONGO_URI"] = 'mongodb://localhost:27017/celery-test-db'
mongo = PyMongo(app)
##############################
# CELERY ARGUMENTS
##############################
app.config['CELERY_BROKER_URL'] = 'amqp://localhost//'
app.config['CELERY_RESULT_BACKEND'] = 'mongodb://localhost:27017/celery-test-db'
app.config['CELERY_RESULT_BACKEND'] = 'mongodb'
app.config['CELERY_MONGODB_BACKEND_SETTINGS'] = {
"host": "localhost",
"port": 27017,
"database": "celery-test-db",
"taskmeta_collection": "celery_jobs",
}
app.config['CELERY_TASK_SERIALIZER'] = 'json'
celery = Celery('task',broker='mongodb://localhost:27017/jobs')
celery = make_celery(app)
LOCK_EXPIRE = 60 * 2 # Lock expires in 2 minutes
@contextmanager
def memcache_lock(lock_id, oid):
timeout_at = monotonic() + LOCK_EXPIRE - 3
# cache.add fails if the key already exists
status = cache.add(lock_id, oid, LOCK_EXPIRE)
try:
yield status
finally:
# memcache delete is very slow, but we have to use it to take
# advantage of using add() for atomic locking
if monotonic() < timeout_at and status:
# don't release the lock if we exceeded the timeout
# to lessen the chance of releasing an expired lock
# owned by someone else
# also don't release the lock if we didn't acquire it
cache.delete(lock_id)
@celery.task(bind=True, name='app.myTask1')
def myTask1(self):
self.update_state(state='IN TASK')
lock_id = self.name
with memcache_lock(lock_id, self.app.oid) as acquired:
if acquired:
# do work if we got the lock
print('acquired is {}'.format(acquired))
self.update_state(state='DOING WORK')
time.sleep(90)
return 'result'
# otherwise, the lock was already in use
raise self.retry(countdown=60) # redeliver message to the queue, so the work can be done later
@celery.task(bind=True, name='app.myTask2')
def myTask2(self):
print('you are in task2')
self.update_state(state='STARTING')
time.sleep(120)
print('task2 done')
@app.route('/', methods=['GET', 'POST'])
def index():
return render_template('index.html')
@app.route('/task1', methods=['GET', 'POST'])
def task1():
print('running task1')
result = myTask1.delay()
# get async task id
taskResult = AsyncResult(result.task_id)
# push async taskid into db collection job_task_id
mongo.db.job_task_id.insert({'taskid': str(taskResult), 'TaskName': 'task1'})
return render_template('task1.html')
@app.route('/task2', methods=['GET', 'POST'])
def task2():
print('running task2')
result = myTask2.delay()
# get async task id
taskResult = AsyncResult(result.task_id)
# push async taskid into db collection job_task_id
mongo.db.job_task_id.insert({'taskid': str(taskResult), 'TaskName': 'task2'})
return render_template('task2.html')
@app.route('/status', methods=['GET', 'POST'])
def status():
taskid_list = []
task_state_list = []
TaskName_list = []
allAsyncData = mongo.db.job_task_id.find()
for doc in allAsyncData:
try:
taskid_list.append(doc['taskid'])
except:
print('error with db conneciton in asyncJobStatus')
TaskName_list.append(doc['TaskName'])
# PASS TASK ID TO ASYNC RESULT TO GET TASK RESULT FOR THAT SPECIFIC TASK
for item in taskid_list:
try:
task_state_list.append(myTask1.AsyncResult(item).state)
except:
task_state_list.append('UNKNOWN')
return render_template('status.html', data_list=zip(task_state_list, TaskName_list))
最终工作代码
from flask import (Flask, render_template, flash, redirect,
url_for, session, logging, request, g, render_template_string, jsonify)
from flask_caching import Cache
from contextlib import contextmanager
from celery import Celery
from Flask_celery import make_celery
from celery.result import AsyncResult
from celery.utils.log import get_task_logger
from celery.five import monotonic
from flask_pymongo import PyMongo
from hashlib import md5
import pymongo
import time
import redis
from flask_redis import FlaskRedis
app = Flask(__name__)
# ADDING REDIS
redis_store = FlaskRedis(app)
# POINTING CACHE_TYPE TO REDIS
cache = Cache(app, config={'CACHE_TYPE': 'redis'})
app.config['SECRET_KEY']= 'super secret key for me123456789987654321'
######################
# MONGODB SETUP
#####################
app.config['MONGO_HOST'] = 'localhost'
app.config['MONGO_DBNAME'] = 'celery-test-db'
app.config["MONGO_URI"] = 'mongodb://localhost:27017/celery-test-db'
mongo = PyMongo(app)
##############################
# CELERY ARGUMENTS
##############################
# CELERY USING REDIS
app.config['CELERY_BROKER_URL'] = 'redis://localhost:6379/0'
app.config['CELERY_RESULT_BACKEND'] = 'mongodb://localhost:27017/celery-test-db'
app.config['CELERY_RESULT_BACKEND'] = 'mongodb'
app.config['CELERY_MONGODB_BACKEND_SETTINGS'] = {
"host": "localhost",
"port": 27017,
"database": "celery-test-db",
"taskmeta_collection": "celery_jobs",
}
app.config['CELERY_TASK_SERIALIZER'] = 'json'
celery = Celery('task',broker='mongodb://localhost:27017/jobs')
celery = make_celery(app)
LOCK_EXPIRE = 60 * 2 # Lock expires in 2 minutes
@contextmanager
def memcache_lock(lock_id, oid):
timeout_at = monotonic() + LOCK_EXPIRE - 3
print('in memcache_lock and timeout_at is {}'.format(timeout_at))
# cache.add fails if the key already exists
status = cache.add(lock_id, oid, LOCK_EXPIRE)
try:
yield status
print('memcache_lock and status is {}'.format(status))
finally:
# memcache delete is very slow, but we have to use it to take
# advantage of using add() for atomic locking
if monotonic() < timeout_at and status:
# don't release the lock if we exceeded the timeout
# to lessen the chance of releasing an expired lock
# owned by someone else
# also don't release the lock if we didn't acquire it
cache.delete(lock_id)
@celery.task(bind=True, name='app.myTask1')
def myTask1(self):
self.update_state(state='IN TASK')
print('dir is {} '.format(dir(self)))
lock_id = self.name
print('lock_id is {}'.format(lock_id))
with memcache_lock(lock_id, self.app.oid) as acquired:
print('in memcache_lock and lock_id is {} self.app.oid is {} and acquired is {}'.format(lock_id, self.app.oid, acquired))
if acquired:
# do work if we got the lock
print('acquired is {}'.format(acquired))
self.update_state(state='DOING WORK')
time.sleep(90)
return 'result'
# otherwise, the lock was already in use
raise self.retry(countdown=60) # redeliver message to the queue, so the work can be done later
@celery.task(bind=True, name='app.myTask2')
def myTask2(self):
print('you are in task2')
self.update_state(state='STARTING')
time.sleep(120)
print('task2 done')
@app.route('/', methods=['GET', 'POST'])
def index():
return render_template('index.html')
@app.route('/task1', methods=['GET', 'POST'])
def task1():
print('running task1')
result = myTask1.delay()
# get async task id
taskResult = AsyncResult(result.task_id)
# push async taskid into db collection job_task_id
mongo.db.job_task_id.insert({'taskid': str(taskResult), 'TaskName': 'myTask1'})
return render_template('task1.html')
@app.route('/task2', methods=['GET', 'POST'])
def task2():
print('running task2')
result = myTask2.delay()
# get async task id
taskResult = AsyncResult(result.task_id)
# push async taskid into db collection job_task_id
mongo.db.job_task_id.insert({'taskid': str(taskResult), 'TaskName': 'task2'})
return render_template('task2.html')
@app.route('/status', methods=['GET', 'POST'])
def status():
taskid_list = []
task_state_list = []
TaskName_list = []
allAsyncData = mongo.db.job_task_id.find()
for doc in allAsyncData:
try:
taskid_list.append(doc['taskid'])
except:
print('error with db conneciton in asyncJobStatus')
TaskName_list.append(doc['TaskName'])
# PASS TASK ID TO ASYNC RESULT TO GET TASK RESULT FOR THAT SPECIFIC TASK
for item in taskid_list:
try:
task_state_list.append(myTask1.AsyncResult(item).state)
except:
task_state_list.append('UNKNOWN')
return render_template('status.html', data_list=zip(task_state_list, TaskName_list))
if __name__ == '__main__':
app.secret_key = 'super secret key for me123456789987654321'
app.run(port=1234, host='localhost')
这里还有一张截图，你可以看到我运行了
myTask1两次和myTask2一次。现在我已经得到了myTask1的期望行为。现在，如果另一个工作人员试图接手它，myTask1将由单个工作人员运行，并根据我定义的重试次数进行重试。
self是一个字符串，而self.cache不存在。这只是一个猜测，但也许Cache.add应该是一个实例，所以像Cache().add这样的形式？因为当调用add函数时，第一个参数可能是 self，例如def add(self, lock_id, oid, lock_expire):，所以根据你目前的代码，self 就是 lock_id？ - jmunschmyTask1时，它只会被排队，直到锁定完成才会运行。 - Max PowersmyTask1没有在使用，那么我希望它能够运行。如果另一个工作者正在使用myTask1，理想情况下我希望任务被排队，但在锁定被移除之前不要运行。希望这样更清楚了。 - Max Powers