You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When running the container on EKS with and configuration it to use EKS Pod Identity, it uses the underlying node's identity and credentials instead of the credentials provided by EKS Pod Identity, resulting in access denied errors.
Expected Behavior
When I configure the pod to use EKS Pod Identity, I expect it to use the credentials provided by EKS Pod Identity, not the host's credentials.
Steps to Reproduce the Bug
Configure and deploy a pod on EKS according to the documentation, with EKS Pod Identity also deployed, and the correct permissions set up.
Environment Details
Version of the S3 container used: DockerHub
Version of NGINX Open Source or NGINX Plus: OSS
Version of NGINX JavaScript: no idea
Target deployment platforms: EKS
S3 backend implementation: AWS
Authentication method: EKS Pod Identity
Additional Context
After investigation, we found ou where the problem lies. The nginx.conf file lists the environment variables that are passed the the worker threads. Just by adding the AWS_CONTAINER_AUTHORIZATION_TOKEN_FILE environment variable to the list of environment variables passed to the worker threads, it works.
TL;DR: adding env AWS_CONTAINER_AUTHORIZATION_TOKEN_FILE; to nginx.conf fixes the issue.
The text was updated successfully, but these errors were encountered:
Bug Overview
When running the container on EKS with and configuration it to use EKS Pod Identity, it uses the underlying node's identity and credentials instead of the credentials provided by EKS Pod Identity, resulting in access denied errors.
Expected Behavior
When I configure the pod to use EKS Pod Identity, I expect it to use the credentials provided by EKS Pod Identity, not the host's credentials.
Steps to Reproduce the Bug
Configure and deploy a pod on EKS according to the documentation, with EKS Pod Identity also deployed, and the correct permissions set up.
Environment Details
Additional Context
After investigation, we found ou where the problem lies. The
nginx.conf
file lists the environment variables that are passed the the worker threads. Just by adding theAWS_CONTAINER_AUTHORIZATION_TOKEN_FILE
environment variable to the list of environment variables passed to the worker threads, it works.TL;DR: adding
env AWS_CONTAINER_AUTHORIZATION_TOKEN_FILE;
tonginx.conf
fixes the issue.The text was updated successfully, but these errors were encountered: