Azure Auto ML JobConfigurationMaxSizeExceeded error when using a cluster

user_13764121 0
Viewed 171

I am running into the following error when I try to run Automated ML through the studio on a GPU compute cluster:

Azure ML error message

Error: AzureMLCompute job failed. JobConfigurationMaxSizeExceeded: The specified job configuration exceeds the max allowed size of 32768 characters. Please reduce the size of the job's command line arguments and environment settings

The attempted run is on a registered tabulated dataset in filestore and is a simple regression case. Strangely, it works just fine with the CPU compute instance I use for my other pipelines. I have been able to run it a few times using that and wanted to upgrade to a cluster only to be hit by this error. I found online that it could be a case of having the following setting: AZUREML_COMPUTE_USE_COMMON_RUNTIME:false; but I am not sure where to put this in when just running from the web studio.

azure-machine-learning-studio azure-machine-learning-service
2 Answers

This is a known bug. I am following up with product group to see if there any update for this bug. For the workaround you mentioned, it need you to go to the node failing with the JobConfigurationMaxSizeExceeded exception and manually set AZUREML_COMPUTE_USE_COMMON_RUNTIME:false in their Environment JSON field.

The node is as below screenshot.enter image description here

user_95988010

It looks like the bug was fixed. I just ran it on a cluster without changing any of the parameters. Thank you Yutong for the help!

user_137641210
Related