Airflow backfill tasks stuck in No status: scheduler starves backfill runs

TL;DR

If backfill tasks sit in No status while scheduled DAGs run fine, check scheduler throughput, not pools: the backfill runs were starving because the scheduler's per-loop run budget never reached them.

Steps

1. If backfill tasks sit in No status while scheduled DAGs run fine, check scheduler throughput, not pools: the backfill runs were starving because the scheduler's per-loop run budget never reached them. On affected versions, raise maxdagrunsperloopto_schedule, run backfills in a quiet window, or upgrade past the merged fix. Restarting pods or re-adding the DAG does nothing: this is scheduler-side starvation.

When to use

You are seeing this: Root cause, confirmed by maintainers in-thread: the scheduler examines only 20 dag runs per loop, and a nulls_first sort on the left-joined backfill sort ordinal pushed backfill dag runs to the very bottom, so on a deployment with ~100 DAGs they starved... Use this skill when you run into "Airflow backfill tasks stuck in No status: scheduler starves backfill runs".

When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

Versions

No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.

Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.

Source: GitHub issue apache/airflow#58554 (closed, 9 comments): backfill tasks stuck in No status on a busy deployment with plenty of free pool slots. The backfill dag run was created and running, but its task instances never got scheduled. Root cause, confirmed by maintainers in-thread: the scheduler examines only 20 dag runs per loop, and a nulls_first sort on the left-joined backfill sort ordinal pushed backfill dag runs to the very bottom, so on a deployment with ~100 DAGs they starved permanently. Maintainers agreed permanent starvation was a bug, not intended deprioritization, and the fix was merged in PR #72730 (2026-09-16).

Verified resolution: If backfill tasks sit in No status while scheduled DAGs run fine, check scheduler throughput, not pools: the backfill runs were starving because the scheduler's per-loop run budget never reached them. On affected versions, raise maxdagrunsperloopto_schedule, run backfills in a quiet window, or upgrade past the merged fix. Restarting pods or re-adding the DAG does nothing: this is scheduler-side starvation.

Source URL: https://github.com/apache/airflow/issues/58554