Vertex AI Pipeline: The replica workerpool0-0 exited with a non-zero status of 13

[norahul1020]

I'm trying to run a vertex AI pipeline (AutoML) using default setting and required service account permission but every time i run that pipeline, i will get failed with above error. 

same i had run with custom will success but when running with AutoML for Tabular Classification / Regression it will get failed after long process time.

[norahul1020]

@lambdalove kindly have a look

[lambdalove]

Sorry @norahul1020 but I'm not familiar with Vertex AI Pipelines or with the error message.

[norahul1020]

Not a problem still looking for the solutions.

[hadi224]

I am running into the same problem and have not been able to figure out why the error is happening. I also saw a similar post on reddit https://www.reddit.com/r/googlecloud/comments/1114k19/help_for_gcpvertexai_error_code_the_replica/

Please post an update here if you resolve the issue. I will do the same. On the surface it appears to be a permission issue. I keep adding more permissions but I have not been able to move forward. It might be something else.

[mohsin-m]

https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/service-accounts

Make sure to have these permissions set

[gatita]

The replica workerpool0-0 exited with a non-zero status of 1. Termination reason: Error.  - Same problem

[bineli_herve]

Hello, same problem here.

Do you have a solution for this problem yet ?

Best Regards,

[Sobhani]

It is likely that "The replica workerpool0-0 exited with a non-zero status of 13" was last error in the job, but previous errors were logged after that with later timestamps. Click on the view logs button and see what were the other errors in job, other than  "The replica workerpool0-0 exited with a non-zero status of 13". For our case, a dataset was assumed to be created manually to create a table in it, so google api complained as it was not there.