Domantas M
08/11/2022, 1:08 PMdagster._core.errors.DagsterInvalidConfigError: Error in config for job
Error 1: Missing required config entry "io_manager" at path root:resources. Sample config for missing entry: {'io_manager': {'config': {'s3_bucket': '...'}}}
Job decorator looks like this:
@job(
resource_defs={"io_manager": s3_pickle_io_manager, "s3": s3_resource, "params": make_values_resource()},
executor_def=k8s_job_executor,
)
run config:
run_config={
"resources": {
"params": {
"config": {
"script_path": "spark_dagster_test.py",
"cluster_name": "spark_pod_name",
"logs_output": "spark_dagster_test.txt",
}
}
}
}
The problem: I could not define s3 required values in dagit UI before launching run since dagit immediately throws an error from the above during workspace loading phrase.
Combination of make_values_resource()
and io_manager/s3
in resource_defs seems does not work together since individually code works as expected.
Does anybody knows where the problem is hidding and could give some advices to solve it?claire
08/11/2022, 5:59 PMconfig
argument to your job, Dagster assumes that it is the full config (so contains all necessary configuration for all resources). If you want to provide partial config for just a single resource, you can use the configured
API instead.
configured API: https://docs.dagster.io/concepts/configuration/configured
@job(
resource_defs={
"io_manager": s3_pickle_io_manager,
"s3": s3_resource,
"params": make_values_resource().configured(
{
"script_path": "spark_dagster_test.py",
"cluster_name": "spark_pod_name",
"logs_output": "spark_dagster_test.txt",
}
),
},
# executor_def=k8s_job_executor,
)
Domantas M
08/17/2022, 10:44 AM