isaac-florence opened a new issue, #28293:
URL: https://github.com/apache/airflow/issues/28293
### Apache Airflow version
Other Airflow 2 version (please specify below)
### What happened
Airflow 2.4.3, running on Kubernetes with High-Availability Scheduler (2
pods). Database is postgres 10.21
I have a dag that has been running for two months with this timetable:
`timetable=CronTriggerTimetable("40 5 * * *", timezone = 'UTC')` (which has not
changed). It has run reliably every day as expected however yesterday it simply
did not run. I have checked for errors and there were none, and checking the
scheduler logs (on both pods) it appears no evaluation of the dag run was done
in either scheduler instance. in the cli inspecting the next 60 run execution
dates indicates there is no problem with the timetable's interpretation
My suspiscion is that a race condition in the `SELECT ... FOR UPDATE` of the
database may have prevented either scheduler executing the DAG. Is this
possible? Otherwise how can this have happened?
### What you think should happen instead
The dag should have run as it always has previously
### How to reproduce
I wish I knew! I can't work it out.
### Operating System
RedHat Enterprise Linux 8
### Versions of Apache Airflow Providers
2.4.3
### Deployment
Other Docker-based deployment
### Deployment details
High availability schedulers (n=2) on OpenShift Kubernetes. Postgres
database (version 10.21
### Anything else
This has happened once before, just under a month ago, on the same DAG.
### Are you willing to submit PR?
- [ ] Yes I am willing to submit a PR!
### Code of Conduct
- [X] I agree to follow this project's [Code of
Conduct](https://github.com/apache/airflow/blob/main/CODE_OF_CONDUCT.md)
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
