How to fix failure at Run_ML_Pipeline job

naeimeh 0

I'm new to azure ml services. I'm using MLOpsPython repo and added my code there and now i'm trying to run the pipeline but it fails at the Trigger ML Training Pipeline. Can someone help me to trboubleshoot the issue?

Here is part of error:

##[debug] /__w/_temp/azureclitaskscript1726870043448.sh

/bin/bash /__w/_temp/azureclitaskscript1726870043448.sh Model was not registered for this run. ##[debug]Exit code 1 received from tool '/bin/bash' ##[debug]STDIO streams have closed for tool '/bin/bash' ##[debug]task result: Failed ##[error]Script failed with error: Error: The process '/bin/bash' failed with exit code 1 at ExecState._setResult (/__w/_tasks/AzureCLI_46e4be58-730b-4389-8a2f-ea10b3e5e815/1.245.5/node_modules/azure-pipelines-task-lib/toolrunner.js:1219:25) at ExecState.CheckComplete (/__w/_tasks/AzureCLI_46e4be58-730b-4389-8a2f-ea10b3e5e815/1.245.5/node_modules/azure-pipelines-task-lib/toolrunner.js:1202:18) at ChildProcess.<anonymous> (/__w/_tasks/AzureCLI_46e4be58-730b-4389-8a2f-ea10b3e5e815/1.245.5/node_modules/azure-pipelines-task-lib/toolrunner.js:1115:19) at ChildProcess.emit (node:events:513:28) at maybeClose (node:internal/child_process:1100:16) at Socket.<anonymous> (node:internal/child_process:458:11) at Socket.emit (node:events:513:28) at Pipe.<anonymous> (node:net:301:12) ##[debug]Processed: ##vso[task.issue type=error;source=TaskInternal;correlationId=ae073d38-90d6-4917-b14b-b1b158fe8373;]Script failed with error: Error: The process '/bin/bash' failed with exit code 1 at ExecState._setResult (/__w/_tasks/AzureCLI_46e4be58-730b-4389-8a2f-ea10b3e5e815/1.245.5/node_modules/azure-pipelines-task-lib/toolrunner.js:1219:25) at ExecState.CheckComplete (/__w/_tasks/AzureCLI_46e4be58-730b-4389-8a2f-ea10b3e5e815/1.245.5/node_modules/azure-pipelines-task-lib/toolrunner.js:1202:18) at ChildProcess.<anonymous> (/__w/_tasks/AzureCLI_46e4be58-730b-4389-8a2f-ea10b3e5e815/1.245.5/node_modules/azure-pipelines-task-lib/toolrunner.js:1115:19) at ChildProcess.emit (node:events:513:28) at maybeClose (node:internal/child_process:1100:16) at Socket.<anonymous> (node:internal/child_process:458:11) at Socket.emit (node:events:513:28) at Pipe.<anonymous> (node:net:301:12) User's image

romungi-MSFT 46,476 Reputation points Microsoft Employee

2024-09-23T06:20:49.49+00:00

@naeimeh The message from the error indicates that your model is not ready or not registered yet and the subsequent run failed since the model is not available yet. See this message:

Model was not registered for this run

I believe you are using this repo for setting up. This repo is mostly using v1 SDK and is not updated since some time, so I think the issue might be that the setup needs an update but the repo went out of date.

I would recommend to lookup the automation documents from Azure ML documentation and try to setup a new pipeline with v2 SDK. You can also use the azureml-examples repo on latest examples for SDK and pipelines that are easy to setup using Jupyter notebooks. Thanks!!
naeimeh 0 Reputation points

2024-09-23T09:22:02.26+00:00

Thank you so much for the suggestion.

Yeah, I'm using this repo and I got that the model is not registered but I can not find why.

Apart from that I'm using Automl from flaml package for training the model. I have only one day to make the pipeline work. What do you think it is the best approach?
romungi-MSFT 46,476 Reputation points Microsoft Employee

2024-09-23T10:12:59.9966667+00:00

@naeimeh I would suggest to try automl pipelines from SDK v2 from here. They should not take much time to run and setup. You can then try updating any of the sample notebooks with your data and run a new pipeline.
romungi-MSFT 46,476 Reputation points Microsoft Employee

2024-09-26T15:22:25.51+00:00

@naeimeh I am curious to check if you are able to get the pipeline to work as you intended. Thanks!!

Share via

How to fix failure at Run_ML_Pipeline job

Your answer