delete_expired method fails for large queryset.

[pyarun]

We process and save results for around 1M task everyday. I see celery.backend_cleanup: 0 4 * * * (m/h/dM/MY/d) UTC in Periodic tasks and it runs everyday. but records are not deleted.

When i ran TaskResult.objects.delete_expired(24*3600) from django shell, shell was killed.

Should we refactor delete_expired to delete objects in chunks?

[auvipy]

we got a PR related to this issue, can you check?

[saz]

FYI: we're having the same issue (currently approx. 93 million rows) and with my patch, the job runs without getting killed. Cleaning up took a few minutes, but as the job will be executed again daily, this won't do any harm.