Is there a way to get failed child tasks to re-run without re-running each child task individually? For example:

import prefect
from prefect import Flow, task
from prefect.engine.results import PrefectResult

from random import random


@task
def make_data():
    return list(range(5))


@task
def randomly_raise(x):
    x = random()
    if x <= 0.5:
        raise Exception()


with Flow("test retries", result=PrefectResult()) as flow:
    data = make_data()
    randomly_raise.map(data)

When the above runs, it creates 5 mapped tasks, each of which randomly fail with 50% chance. When there are failed runs, if I click re-run on the failed tasks, I can get them to individually re-run & succeed. But, re-running on the entire flow never attempts the mapped tasks again. Here’s what’s logged

June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise': task is mapped, but run will proceed so children are generated."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise': Handling state change from Mapped to Mapped"
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[0]': Starting task run..."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise[0]': task is already finished."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[0]': finished task run for task with final state: 'Success'"
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[1]': Starting task run..."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise[1]': task is already finished."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[1]': finished task run for task with final state: 'Success'"
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[2]': Starting task run..."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise[2]': task is already finished."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[2]': finished task run for task with final state: 'Failed'"
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[3]': Starting task run..."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise[3]': task is already finished."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[3]': finished task run for task with final state: 'Failed'"
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[4]': Starting task run..."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,DEBUG,"Task 'randomly_raise[4]': task is already finished."
June 12th 2020,9:22:35am ,prefect.CloudTaskRunner,INFO,"Task 'randomly_raise[4]': finished task run for task with final state: 'Failed'"

Is there a way to re-run a whole group of mapped tasks?

Hey @Howard Cornwell I believe restarting from failed will only rerun tasks which are not successful in order to avoid duplicate work. Starting another run is one way of rerunning the whole group

Thanks, perhaps I should clarify the last question: is there a way to re-run all the failed child tasks of a mapped task?

Thanks for the clarification! Looking into it

Thanks, good to know I’m not going mad!

@Howard Cornwell let me check on that. The issue was UI related and not actually anything in the core code itself 🙂 will let you know when I find out

Thanks for the update