failed to create new S3 client: Endpoint: $(kfp-artifact-storage-endpoint) does not follow ip address or domain name standards

[TranThanh96]

Describe the bug
I deployed RDS + s3 option.
Follow the doc to verify s3: https://awslabs.github.io/kubeflow-manifests/docs/deployment/rds-s3/guide/#42-verify-s3

when running the sample pipeline "[Demo] XGBoost - Iterative model training", a bug occurred: failed to create new S3 client: Endpoint: $(kfp-artifact-storage-endpoint) does not follow ip address or domain name standards

all pods are still running.

How can I fix that?

[ryansteakley]

@TranThanh96 Did the fix, I sent you on slack work for you? We have fixed this in the latest commit.

[LeoBley]

@ryansteakley Can I get that fix to or has it been released already?

[TranThanh96]

@ryansteakley It works. Thank you

@LeoBley edit file kubeflow-manifests/awsconfigs/apps/pipeline/s3/config

change endpoint: $(kfp-artifact-storage-endpoint) to endpoint: s3.amazonaws.com
then apply by running install loop again

[LeoBley]

Thanks @TranThanh96 .

That fixed it, and I also finally figured out I needed to configure the run parameters with:
pipeline-root = minio://bucketName/ where bucketName is the value we entered in the awsconfigs/apps/pipeline/s3/params.env file

[surajkota]

Hi everyone, thanks for reporting this issue. We have fixed this and released a patch version https://github.com/awslabs/kubeflow-manifests/releases/tag/v1.5.1-aws-b1.0.2. Please reopen this issue, if you are still facing it