The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

Next hurdle I've run into running dagster on ECS is that my repo is loaded and displayed in Dagit correctly, but when I run the pipeline, I get:
```[EcsRunLauncher] Launching run in task arn:aws:ecs:us-east-1:726618033885:task/data-dagster-prod/2bc319b2cb104976836aacf81a6300d8 on cluster arn:aws:ecs:us-east-1:726618033885:cluster/data-dagster-prod
Could not load pipeline definition.
FileNotFoundError: [Errno 2] No such file or directory: 'recommendations/repo.py'```


I saw in <https://docs.dagster.io/deployment/guides/docker|the guide>:
&gt; This launcher will start each run in a new container, using whatever image that you set in the `DAGSTER_CURRENT_IMAGE` environment variable in your pipeline container (which will usually be the same image as the pipeline container itself)
And I do have `DAGSTER_CURRENT_IMAGE: "$REGISTRY_URL/deploy_ecs/pipelines"` in my pipelines definition.

My understanding is that since the DAG is displayed in dagit, that the pipelines contains has successfully served `$DAGSTER_HOME/recommendations/repo.py`, but, seemingly it isn't found at execution time.  Does the run start with `$DAGSTER_HOME` has its working directory?

I see in the actual task:
```\"repository_origin\": {\"__class__\": \"RepositoryPythonOrigin\", \"code_pointer\": {\"__class__\": \"FileCodePointer\", \"fn_name\": \"repository\", \"python_file\": \"recommendations/repo.py\", \"working_directory\": \"/opt/dagster/dagster_home\"}, ```
Which tells me that the working directory is `/opt/dagster/dagster_home/` and that it is trying to load `recommendations/repo.py`.  That should just work, as here it is from within the `pipelines` container:
```root@63629cdf2b98:/opt/dagster/dagster_home# pwd
/opt/dagster/dagster_home
root@63629cdf2b98:/opt/dagster/dagster_home# ls -al recommendations/repo.py
-rw-r--r-- 1 root root 12523 Sep 23 14:57 recommendations/repo.py```
What am I missing here?

I figured this out... the Run task wasn't using the tagged docker image.