Don't fail DagRun when leaf mapped_task is SKIPPED #25995
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Member
Author
|
/cc @collinmcnulty @ldacey I think this should fix the issue if you want to try it out. |
ephraimbuddy
approved these changes
Aug 26, 2022
ashb
force-pushed
the
fix-mapped-leaf-task-skip-dagrun-deadlock
branch
from
510ded2 to
ab64e0f
Compare
This one was a fun one to track down, and was only a problem when the scheduler "expanded" the first mapped task (making it SKIPPED). - The scheduler looks at `add_one` and marks it as SKIPPED. - `unfinished_tis` contains add_one_1, and `_are_premature_tis` changes the state of `add_one__1`, (which just so happens to be a leaf task), and the check on line 584 essentially gets confused, as no one envisaged the states changing! In a reverse of how this normally plays out, if the mini_scheduler in the LocalTaskJob was disabled then this example DAG would deadlock every time. (Since that mini scheduler only operates on a partial DAG it can't ever change the whole DagRun state.) Co-authored-by: Ephraim Anierobi <splendidzigy24@gmail.com>
uranusjr
force-pushed
the
fix-mapped-leaf-task-skip-dagrun-deadlock
branch
from
940cb22 to
d98ca51
Compare
Member
Author
|
Mypy is confusing me: Can't get a "good" fix, having to add a type-ignore. (I'm getting conflicting behaviour. I can fix it either for all files, or for a single file, but not both) |
ashb
commented
Aug 31, 2022
| ) | ||
|
|
||
| self.trigger_rule = TriggerRule(trigger_rule) | ||
| self.trigger_rule: TriggerRule = TriggerRule(trigger_rule) |
Member
Author
There was a problem hiding this comment.
Mypy really can't work this out? :o
Member
There was a problem hiding this comment.
I think it has something to do with either (or a combination of) Enum having a custom metaclass, TriggerRule does multi-inheritance, and one of the parents is a built-in.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This one was a fun one to track down, and was only a problem when the scheduler "expanded" the first mapped task (making it SKIPPED).
add_oneand marks it as SKIPPED.unfinished_tiscontainsadd_one__1, and_are_premature_tischanges the state ofadd_one__1, (which just so happens to be a leaf task), and the check on line 584 essentially gets confused, as no one envisaged the states changing!In a reverse of how this normally plays out, if the mini_scheduler in the LocalTaskJob was disabled then this example DAG would deadlock every time. (Since that mini scheduler only operates on a partial DAG it can't ever change the whole DagRun state.)
Fixes #25200
Ephraim had the idea of the fix, I just optomized it to not check the (possibly) large list everytime but only when a state was changed.
^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named
{pr_number}.significant.rstor{issue_number}.significant.rst, in newsfragments.