So I believe I have everything deployed correctly ...
# deployment-kubernetes
f
So I believe I have everything deployed correctly for local K8 + Celery but my pipeline is not actually running. Am I missing something that actually tells celery to "grab" the job and process?
Copy code
helm upgrade --install dagster-release --namespace dagster dagster/dagster -f values.yaml --set runLauncher.type=CeleryK8sRunLauncher \
--set dagsterDaemon.queuedRunCoordinator.enabled=true \
--set rabbitmq.enabled=true \
--set flower.enabled=true
n
hi Felipe, one thing I might check is whether that run coordinator job mentioned in your first screenshot launched successful and if there are any logs from the associated pod
plus1 1
j
The run worker pod name is logged in your first screenshot, the commands would be
Copy code
kubectl describe pod dagster-run-<ID>
kubectl logs dagster-run-<ID>
f
Thanks, using describe/logs allowed me to see part of my issue. I am only running this locally at the moment and needed to change the image_pull_policy and the job_namespace
Copy code
execution:
  celery-k8s:
    config:
      image_pull_policy: IfNotPresent
      job_namespace: dagster
j
Glad you were able to work this out, it’s working as expected now?
We also are tracking a few issues similar to what you ran into here where we don’t expose k8s error info in Dagit. Feel free to contribute any thoughts! https://github.com/dagster-io/dagster/issues/2856
f
Well, not exactly. I need to read through how _env_config_maps_ work. I am getting issues like below
Copy code
dagster.core.errors.DagsterInvalidConfigError: Errors whilst loading configuration for {'instance_config_map': Field(<dagster.config.source.StringSourceType object at 0x7f97a3d385d0>, default=@, is_required=True), 'postgres_password_secret': Field ......
j
Is there more to the error message? Where is this error surfaced?
f
I was not pointing to a valid configmap. I pointed to a valid one and got celery to actually kick things off.
j
🎉
🎉 1