I should have been clearer, in my case the tasks were marked as shutdown. This was reflected in the *state* column in the *task_instance* table.
On Sun, Mar 5, 2017 at 2:18 PM Bolke de Bruin <[email protected]> wrote: > Can you provide a bit more details on “SUCCEEDED” vs “FAILURE”? We use the > db as a state keeper. And only the Task itself can mark SUCCESS or FAILED. > Som I am wondering were did you see those states? > > B > > > On 2 Mar 2017, at 15:27, twinkle <[email protected]> wrote: > > > > Hi, > > > > We plan to use Airflow along with Celery as the backend. > > Today within a DAG run, despite showing some of the tasks in a DAG as > > successful, Airflow was not scheduling the next potential tasks in it. > > Looking at the Celery Flower, following exceptions were observed: > > > > Traceback (most recent call last): > > File > > > "/home/allocation/.pyenv/versions/2.7.12/lib/python2.7/site-packages/celery/app/trace.py", > > line 367, in trace_task > > R = retval = fun(*args, **kwargs) > > File > > > "/home/allocation/.pyenv/versions/2.7.12/lib/python2.7/site-packages/celery/app/trace.py", > > line 622, in __protected_call__ > > return self.run(*args, **kwargs) > > File > > > "/home/allocation/.pyenv/versions/2.7.12/lib/python2.7/site-packages/airflow/executors/celery_executor.py", > > line 45, in execute_command > > raise AirflowException('Celery command failed') > > AirflowException: Celery command failed > > > > There has been no failure logs at the Airflow side, and it has marked the > > task as Succeeded. > > > > Looking at the meta data table, i found the state of the task as FAILURE. > > It seems like some of the link is broken, as to some extent Airflow > > realises the failure, due to which it stopped scheduling the tasks > further, > > but it is not complete, as the UI showed different state. > > > > Has anyone else experienced it? > > > > Regards, > > Twinkle > > -- Sergei
