Fix DAG.cli() crashing on dags pause and unpause - #72109
Open
Eason09053360 wants to merge 2 commits into
Open
Conversation
A Dag file run as a script goes through DAG.cli(), whose parser drops --dag-id and instead dispatches every subcommand with the Dag object as a second positional argument. The pause and unpause handlers never accepted it, so Dag authors got a TypeError instead of the command. The handlers are only reachable this way from a Dag file, which is why the regular `airflow dags pause` path has always worked and no test covered the difference.
Eason09053360
requested review from
bugraoz93,
dheerajturaga,
henry3260 and
potiuk
as code owners
August 26, 2026 16:32
ColtenOuO
reviewed
Aug 26, 2026
ColtenOuO
left a comment
ColtenOuO
left a comment
Contributor
There was a problem hiding this comment.
Great work!
However, once the issue addressed in this PR is resolved, it seems to introduce another problem.
The Dag-scoped parser drops --dag-id but keeps --treat-dag-id-as-regex, so the flag survives into a context where the user has no pattern left to supply. The Dag's own id was then read back as a pattern, and dag_ids may contain dots, so an unanchored match could pause unrelated Dags. Normalising the flag next to the dag_id it guards covers both of its readers, including the confirmation prompt, rather than guarding each reader in turn and leaving the next one to be found later.
ColtenOuO
approved these changes
Aug 27, 2026
ColtenOuO
left a comment
ColtenOuO
left a comment
Contributor
There was a problem hiding this comment.
Thanks, Looks great!
Let's wait for maintainer take a more look.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
DAG.cli()dispatches every subcommand with the Dag object as a secondpositional argument, because its parser drops
--dag-id. Thedags pauseand
dags unpausehandlers never accepted it, so running a Dag file as ascript raised
TypeError: dag_pause() takes 1 positional argument but 2 were giveninstead of the command.The regular
airflow dags pause --dag-id foopath has always worked, whichis why this went unnoticed — the handlers are only reachable with two
arguments from a Dag file.
Only
pauseandunpausewere affected.DAG_CLI_DICTresolves to exactlysix subcommands —
dags list-runs/pause/unpause/testandtasks list/test—and the other four already accepted the Dag argument.
The filter at
cli_config.py:2362also listsbackfillandrun, but neithername exists in
DAGS_COMMANDS/TASKS_COMMANDSany more, so they match nothing.Harmless today, and left alone here to keep this diff focused — but it is the
same fragility that hid this bug: two layers wired together by string matching,
with nothing enforcing that the names or the call signatures line up.
Was generative AI tooling used to co-author this PR?
Generated-by: Claude Code (Opus 5) following the guidelines