CommandRunner failing to export Access Keys

Nobody has claimed this yet.

I've been trying to run a kubectl command via CommandRunner and it appears the AWS credentials I'm calling aren't being recognized:

The script updates the ec2, uninstalls cli, reinstalls with updated version, installs kubectl, export the AWS keys, configures aws eks, then runs kubectl.
First I tried just exporting the aws keys:

"Fn::Sub": "sudo yum update -y && sudo yum remove awscli -y && curl 'https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip' -o 'awscliv2.zip' && unzip awscliv2.zip && sudo ./aws/install --bin-dir /usr/local/bin --install-dir /usr/local/aws-cli --update && curl -o /home/ec2-user/kubectl https://s3.us-west-2.amazonaws.com/amazon-eks/1.29.6/2024-07-12/bin/linux/amd64/kubectl && sudo chmod +x ./home/ec2-user/kubectl && mkdir -p /home/ec2-user/bin && cp ./home/ec2-user/kubectl /home/ec2-user/bin/kubectl && echo 'copied kubectl to home/ec2-user/bin' && cp ./home/ec2-user/kubectl /usr/local/bin/kubectl && echo 'copied kubectl to usr/local/bin' && aws s3 cp s3://xxx/yamls/cluster-autoscaler.yaml /home/ec2-user/cluster-autoscaler.yaml && echo 'downloaded yaml' && export AWS_ACCESS_KEY_ID='XXX' && export AWS_SECRET_ACCESS_KEY='XXX' && export AWS_SESSION_TOKEN='XXX' && aws eks --region us-east-2 update-kubeconfig --name test-eks-cluster && kubectl apply -f /home/ec2-user/cluster-autoscaler.yaml && echo success > /command-output.txt"

I then tried creating the .aws/credentials and .aws/config files:
"sudo yum update -y && sudo yum remove awscli -y && curl 'https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip' -o 'awscliv2.zip' && unzip awscliv2.zip && sudo ./aws/install --bin-dir /usr/local/bin --install-dir /usr/local/aws-cli --update && curl -o /home/ec2-user/kubectl https://s3.us-west-2.amazonaws.com/amazon-eks/1.29.6/2024-07-12/bin/linux/amd64/kubectl && sudo chmod +x ./home/ec2-user/kubectl && mkdir -p /home/ec2-user/bin && cp ./home/ec2-user/kubectl /home/ec2-user/bin/kubectl && echo 'copied kubectl to home/ec2-user/bin' && cp ./home/ec2-user/kubectl /usr/local/bin/kubectl && echo 'copied kubectl to usr/local/bin' && aws s3 cp s3://xxxxx/yamls/cluster-autoscaler.yaml /home/ec2-user/cluster-autoscaler.yaml && echo 'downloaded yaml' && mkdir -m 777 /home/ec2-user/.aws/ && printf '[default]\naws_access_key_id=XXX\naws_secret_access_key=XXX\naws_session_token=XXX' >/home/ec2-user/.aws/credentials && printf '[default]\nregion=us-east-1\noutput=json' > /home/ec2-user/.aws/config && aws eks --region us-east-2 update-kubeconfig --name test-eks-cluster && kubectl apply -f /home/ec2-user/cluster-autoscaler.yaml && echo success > /command-output.txt"

With both of these I get the same error:
error: error validating "/home/ec2-user/cluster-autoscaler.yaml": error validating data: failed to download openapi: Get "http://localhost:8080/openapi/v2?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused; if you choose to ignore these errors, turn validation off with --validate=false

If I remote into the ec2, run the export of the aws keys + aws eks + kubectl command, everything works fine.

Neither of these work. Anyone have the same issue?

used both config and credentials

Start with the CommandRunner shell sequence in the issue and inspect how /home/ec2-user/.aws/credentials, /home/ec2-user/.aws/config, and the kubeconfig used by aws eks are handled. Compare the remote-login environment with the command environment; done means kubectl connects to the EKS cluster instead of falling back to localhost:8080.

Written by the indexing model from the issue text.