Image PullBack Error in EKS-A 23.2 with containerd (connection reset by peer) #10205
Description
What happened: We are encountering an image pull-back error in EKS Anywhere version 23.2, specifically on nodes using containerd as the container runtime. The issue is affecting pod scheduling and service availability.
After multiple retires image was succes
What you expected to happen:
How to reproduce it (as minimally and precisely as possible):
Deploy a pod using a public/private image.
Observe pod status in kubectl get pods.
Check events using kubectl describe pod .
Or directly log in to node/vm and use ctr to pull images.
Anything else we need to know?:
Verified image accessibility from other environments.
Checked containerd logs (journalctl -u containerd).
Ensured registry credentials are correctly configured.
Validated network access to image registry.
After multiple retries, the image was eventually pulled successfully.
The same image was tested with containerd.io 1.7.27 and did not exhibit any pull issues
Cluster Details:
EKS-A Version: 23.2
Container Runtime: containerd
Affected Nodes: All nodes using containerd
Cluster Type: Nutanix
Expected Behavior:
Images should be pulled reliably without requiring multiple retries