amichai07 commented on a change in pull request #7402: [AIRFLOW-3607] fix
scheduler bug related to task concurrency and depends on past
URL: https://github.com/apache/airflow/pull/7402#discussion_r381985351
##########
File path: airflow/models/dagrun.py
##########
@@ -284,15 +284,27 @@ def update_state(self, session=None):
none_depends_on_past = all(not t.task.depends_on_past for t in
unfinished_tasks)
none_task_concurrency = all(t.task.task_concurrency is None
for t in unfinished_tasks)
- # small speed up
- if unfinished_tasks and none_depends_on_past and none_task_concurrency:
+ if unfinished_tasks:
scheduleable_tasks = [ut for ut in unfinished_tasks if ut.state in
SCHEDULEABLE_STATES]
+ if none_depends_on_past and none_task_concurrency:
+ # small speed up
+ self.log.debug(
+ "number of scheduleable tasks for %s: %s task(s)",
+ self, len(scheduleable_tasks))
+ ready_tis, changed_tis =
self._get_ready_tis(scheduleable_tasks, finished_tasks, session)
+ self.log.debug("ready tis length for %s: %s task(s)", self,
len(ready_tis))
+ are_runnable_tasks = ready_tis or self._are_premature_tis(
+ unfinished_tasks, finished_tasks, session) or changed_tis
+ else:
+ # slow path
+ for ti in scheduleable_tasks:
+ if ti.are_dependencies_met(
Review comment:
why not use `get_ready_tis`?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
