问题
Suppose i have this task definition:
def some_other_foo(input)
raise Exception('This is not handled!')
return input
@app.task(
bind=True,
max_retries=5,
soft_time_limit=20)
def some_foo(self, someInput={}):
response=""
try:
response = some_other_foo(someInput)
except Exception as exc:
self.retry(countdown=5, exc=exc)
response="error"
return response
I have a problem that exception is not handled in some_foo, I get error instead of response="error", task is crashed and i get Traceback that indicates an Exception was raised.
Is it possible to return regular response, but to set celery task as failed, so result in flower will be failed ?
I am using:
Celery 4.1
AMPQ as broker
Celery Flower as monitoring
回答1:
Try \ except
works fine. Your task will always be unsuccessful because you called self.retry
before return
. Let's make a little test:
from celery import Celery
app = Celery(name_app,broker_settings_etc....)
def some_other_foo(value):
raise Exception('This is not handled!')
@app.task(
bind=True,
max_retries=5,
soft_time_limit=20)
def some_foo(self):
response = ""
try:
response = some_other_foo('test')
except Exception as exc:
self.retry(countdown=5, exc=exc)
response = "error"
return response
Run celery app and call our task. You will see in celery logs something like this:
3fb-81de-e4149fa88f4c] retry: Retry in 5s: Exception('This is not handled!',)
[2017-08-18 15:50:34,160: INFO/MainProcess] Received task: tasks.some_foo[b656731b-c85d-43fb-81de-e4149fa88f4c] eta:[2017-08-18 12:50:39.156912+00:00]
[2017-08-18 15:50:34,161: INFO/MainProcess] Task tasks.some_foo[b656731b-c85d-43fb-81de-e4149fa88f4c] retry: Retry in 5s: Exception('This is not handled!',)
[2017-08-18 15:50:39,511: ERROR/MainProcess] Task tasks.some_foo[b656731b-c85d-43fb-81de-e4149fa88f4c] raised unexpected: Exception('This is not handled!',)
Traceback (most recent call last):
# trace here...
Exception: This is not handled!
How it works. You set for the task max_retries=5
. When you called self.retry(countdown=5, exc=exc)
Celery interrupts processing of task and try to restart task with countdown
(in our case = 5). After 5 attempts(max_retries
) Celery won't rerun task.
Now, let's change our try \ except
block to:
try:
response = some_other_foo('test')
except Exception:
print 'handled'
response = "bad response"
Restart Celery and run our task. Let's check log:
[2017-08-18 15:58:41,893: INFO/MainProcess] Received task: tasks.some_foo[1437e7ce-1c69-4042-824b-5602f486c025]
[2017-08-18 15:58:41,895: WARNING/Worker-3] handled
[2017-08-18 15:58:41,896: INFO/MainProcess] Task tasks.some_foo[1437e7ce-1c69-4042-824b-5602f486c025] succeeded in 0.00186271299026s: 'bad response'
As you can see handler works fine.
So, summarize. If you call self.retry
, Celery will interrupt processing of task and try to restart a current task.
来源:https://stackoverflow.com/questions/45611122/celery-exception-handling