potiuk commented on PR #31001:
URL: https://github.com/apache/airflow/pull/31001#issuecomment-1538769106
I believe, this failing test shows an actual problem (i.e. if we merge it,
`docker-compose` of ours will stop working).
This tests starts a docker-compose of ours and triggers a dag there, I
don't **think** the errors you see see there are related.
It looks like that for whatever reason the docker-compose of ours things
that the airflow worker is unhealthy:
```
$ docker ps
CONTAINER ID IMAGE
COMMAND CREATED STATUS
PORTS NAMES
4255c86396b6
ghcr.io/apache/airflow/main/prod/python3.7:5efaf8bdf0bd84f05049479cf387425000f4771e
"/usr/bin/dumb-init …" 5 minutes ago Up 5 minutes (healthy) 8080/tcp
quick-start_airflow-scheduler_1
b8a7cb3c9429
ghcr.io/apache/airflow/main/prod/python3.7:5efaf8bdf0bd84f05049479cf387425000f4771e
"/usr/bin/dumb-init …" 5 minutes ago Up 5 minutes (unhealthy) 8080/tcp
quick-start_airflow-worker_1
011a97dfdeff
ghcr.io/apache/airflow/main/prod/python3.7:5efaf8bdf0bd84f05049479cf387425000f4771e
"/usr/bin/dumb-init …" 5 minutes ago Up 5 minutes (healthy) 8080/tcp
quick-start_airflow-triggerer_1
d3113c6b9d22
ghcr.io/apache/airflow/main/prod/python3.7:5efaf8bdf0bd84f05049479cf387425000f4771e
"/usr/bin/dumb-init …" 5 minutes ago Up 5 minutes (healthy)
0.0.0.0:8080->8080/tcp, :::8080->8080/tcp quick-start_airflow-webserver_1
4cd33310af92 redis:latest
"docker-entrypoint.s…" 5 minutes ago Up 5 minutes
(healthy) 6379/tcp quick-start_redis_1
bc474e8ce123 postgres:13
"docker-entrypoint.s…" 5 minutes ago Up 5 minutes
(healthy) 5432/tcp quick-start_postgres_1
$ docker-compose logs
```
From the logs it seems that the worker starter correctly, So I think it's
the health check that stopped detecting if the worker is running.
And that I think is somehow related with the change. If it is happening
repeatedly in this PR, Likely the health check for the worker in our docker
compose gets broken by this change (or mayne irflow :
https://github.com/apache/airflow/blob/ffe3a68f9ada2d9d35333d6a32eac2b6ac9c70d6/docs/apache-airflow/howto/docker-compose/docker-compose.yaml#L148
This seems related, the change is about celery executor and that's the one
used in our docker-compose.
The way how to test the docker-compose:
* `breeze prod-image build` -> it will build the image locally (ghcr.io/
.....) - you will see it in the ougput
* `breeze testing docker-compose-tests` - should attempt to run the same
tests
You can also manualy start our docker-compose following the
https://airflow.apache.org/docs/apache-airflow/stable/howto/docker-compose/index.html
steps (because this is what the test is actually doing) - just set the set
AIRFLOW_IMAGE_NAME to point to the prod image build and it should reveal
similar issue.
--
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]
