sdkv2 job on local compute fails with registry authentication error

uberj 20

Steps to reproduce:

Run code from: https://learn.microsoft.com/en-us/azure/machine-learning/migrate-to-v2-local-runs?view=azureml-api-2

#import required libraries
from azure.ai.ml import MLClient, command
from azure.ai.ml.entities import Environment
from azure.identity import DefaultAzureCredential

#connect to the workspace
ml_client = MLClient.from_config(DefaultAzureCredential())

# set up pytorch environment
env = Environment(
    image='mcr.microsoft.com/azureml/openmpi3.1.2-ubuntu18.04',
    conda_file='pytorch-env.yml',
    name='pytorch-env'
)

# define the command
command_job = command(
    code='./src',
    command='train.py',
    environment=env,
    compute='local',
)

returned_job = ml_client.jobs.create_or_update(command_job)
returned_job

Expected: Job runs

Actual:

Traceback (most recent call last):
  File "test.py", line 23, in <module>
    returned_job = ml_client.jobs.create_or_update(command_job)
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/core/tracing/decorator.py", line 76, in wrapper_use_tracer
    return func(*args, **kwargs)
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/ai/ml/_telemetry/activity.py", line 337, in wrapper
    return_value = f(*args, **kwargs)
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/ai/ml/operations/_job_operations.py", line 609, in create_or_update
    raise ex
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/ai/ml/operations/_job_operations.py", line 573, in create_or_update
    snapshot_id = start_run_if_local(
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/ai/ml/operations/_local_job_invoker.py", line 415, in start_run_if_local
    cr_helper.get_bootstrapper_binary(bootstrapper_info)
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/ai/ml/operations/_local_job_invoker.py", line 310, in get_bootstrapper_binary
    docker_client = self.get_docker_client(registry)
  File "/home/uberj/anaconda3/envs/mlopspython_ci/lib/python3.8/site-packages/azure/ai/ml/operations/_local_job_invoker.py", line 232, in get_docker_client
    if registry:
RuntimeError: Login to Docker registry 'mcr.microsoft.com' failed. See error message: 'username'

YutongTie-MSFT 50,856 Reputation points

2023-05-25T23:14:25.7133333+00:00
Hello @uberj

Thanks for reaching out to us, it seem you have not set your DefaultAzureCredential() well, could you please double check on that?

In a shell environment, you can set up DefaultAzureCredential() by installing the Azure CLI and logging in with your Azure account. Once you have logged in, you can use the following command to set up the credential**3**:

az login

This command will authenticate you with Azure and create a default Azure credential that can be used.

I hope this helps.

Regards,

Yutong
uberj 20 Reputation points

2023-05-26T14:57:48.5366667+00:00

My client is logged in and the error still happens. If I change "local" to "train-cluster" (which is the name of an existing compute in my workspace), the script works.
uberj 20 Reputation points

2023-05-26T15:07:27.5466667+00:00

My client is logged in and the error still happens. If I change "local" to "train-cluster" (which is the name of an existing compute in my workspace), the script works.
Gianluca Rondinini 15 Reputation points

2023-07-14T07:19:47.3766667+00:00

Same problem here, with the Azure ML SDK V2 I just can't use compute = "local", else I get the following:

"Login to Docker registry 'mcr.microsoft.com' failed. See error message: 'username'

KeyError: 'username'"
Peiqing Zhang 0 Reputation points

2023-07-25T10:57:35.73+00:00

Same problem here, with the Azure ML SDK V2 and compute = "local".
Murli Sivashanmugam 0 Reputation points

2023-08-14T11:35:06.05+00:00

I have the same problem as well. Any solution?
Farbod Taymouri 0 Reputation points

2023-09-07T14:12:11.6333333+00:00

Same problem here, with the Azure ML SDK V2 and compute = "local" it returns KeyError: 'username

Share via

sdkv2 job on local compute fails with registry authentication error

Your answer