By using AWS re:Post, you agree to the AWS re:Post Terms of Use
AWS re:Postre:Post

Connection Resets when connecting to internal registry from EKS

0

Hi All,

We are currently facing an issue in our EKS cluster. When using DIND inside a gitlab runner pod we are getting connection reset when pushing to an instance of artifactory running on prem. The weird thing is that when images are small, we are able to push successfully but the moment they are big we get issues. We have checked our firewalls, our nginx configs on the internal registry and we have also checked for connection time out settings both on AWS and on prem to no luck. Can someone who has faced a similar issue please assist with how we can resolve or further investigate the issue. Thanks in advance.

  • Adeleke Adebowale J EXPERT
    a month ago

    could you check your EKS cluster nodes, especially those hosting the GitLab runner pods, are not experiencing resource constraints such as high CPU or memory usage. Insufficient resources can cause network issues, especially when handling large image transfers.

Topics
Networking & Content DeliveryComputeContainersDevOps
Tags
Amazon VPCHigh Performance ComputeAmazon EC2Amazon Elastic Kubernetes ServiceDevOps
Language
English
rePost-User-8428534
asked a month ago744 views
1 Answer
0

Increase timeout settings across your GitLab runner, Docker daemon, and any network devices involved in the transmission path. Also, ensure your NAT tables and connection tracking settings are optimized for high volumes of network traffic​

https://github.com/cilium/cilium/issues/21853

profile picture
EXPERT
Sedat Salman
answered a month ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions

Relevant content