芹菜任务在没有芹菜节拍的情况下调用自己一个任务成功

Question

我想在当前任务完成后每隔 30 分钟调用一次我的芹菜任务，但有时任务花费的时间比预期的要长，因为任务是从远程服务器下载文件。所以我不想用芹菜节拍。另外，使用自我。我想重试仅适用于发生错误时。这是我的任务：

@shared_task(name="download_big", bind=True, acks_late=true, autoretry_for=(Exception, requests.exceptiosn.RequestException), retry_kwargs={"max_retries": 4, "countdown": 3}):
def download_big(self):
    my_file = session.get('example.com/hello.mp4')
    if my_file.status_code == requests.codes["OK"]:
        open("hello.mp4", "wb").write(my_file.content)
    else:
        self.retry()

更新：

好吧，我将我的结构更改为：

@shared_task(name="download_big", bind=True, acks_late=true, autoretry_for=(Exception, requests.exceptiosn.RequestException), retry_kwargs={"max_retries": 4, "countdown": 3}):
def download_big(url):
    my_file = session.get(url, name)
    if my_file.status_code == requests.codes["OK"]:
        open(name, "wb").write(my_file.content)
    else:
        self.retry()
@shared_task(name="download_all", bind=True, acks_late=true, autoretry_for=(Exception, requests.exceptiosn.RequestException), retry_kwargs={"max_retries": 4, "countdown": 3}):
def download_all(self):
    my_list = [...]  # bunch of urls with names
    jobs = []
    for name, url in my_list:
        jobs.append(download_big.si(url, name))
    group(jobs)()

所以在这种情况下，我必须调用download_all方法而不是download_big，这样我就可以并行下载文件，当所有组任务完成后，需要在 30 分钟后再次调用它自己。

Answer 1

您可以尝试使用一个 chord，它将运行一组任务，当它们完成时，将运行一个可用于重新调度的回调。

例如

from celery import chord
@shared_task(name="download_big", bind=True, acks_late=true, autoretry_for=(Exception, requests.exceptiosn.RequestException), retry_kwargs={"max_retries": 4, "countdown": 3}):
def download_big(url):
    my_file = session.get(url, name)
    if my_file.status_code == requests.codes["OK"]:
        open(name, "wb").write(my_file.content)
    else:
        self.retry()
@shared_task(name="download_all", bind=True, acks_late=true, autoretry_for=(Exception, requests.exceptiosn.RequestException), retry_kwargs={"max_retries": 4, "countdown": 3}):
def download_all(self):
    my_list = [...]  # bunch of urls with names
    jobs = []
    for name, url in my_list:
        jobs.append(download_big.si(url, name))
    # Run the group and reschedule once all tasks complete
    chord(jobs)(download_all.apply_async(countdown=1800))