agent pod in crash loop unless IPv6 disabled

[edlevin6612]

Launched a new EKS 1.30 cluster with a number of managed node groups backed by a variety of EKS Optimized AMIs (Amazon Linux 2, Amazon Linux 2023, x86 and ARM).

Noticed that eks-pod-identity-agent pods are in a crash loop on each node. Log shows:

ind-addr":"169.254.170.23:80","level":"info","msg":"Starting server...","time":"2024-08-15T19:35:24Z"}
{"bind-addr":"localhost:2703","level":"info","msg":"Starting server...","time":"2024-08-15T19:35:24Z"}
{"bind-addr":"[fd00:ec2::23]:80","level":"info","msg":"Starting server...","time":"2024-08-15T19:35:24Z"}
{"bind-addr":"[fd00:ec2::23]:80","level":"fatal","msg":"Unable to start server: listen tcp [fd00:ec2::23]:80: bind: cannot assign requested address","time":"2024-08-15T19:35:24Z"}
2024/08/15 19:35:24 Now listening for interrupts
2024/08/15 19:35:24 running command: exit status 1

I worked around the problem by disabling IPv6 on the add-on per https://docs.aws.amazon.com/eks/latest/userguide/pod-id-agent-config-ipv6.html

But curious why the agent is broken out of the box for ULA addresses

Thank you

[kmala]

can you look at consideration https://docs.aws.amazon.com/eks/latest/userguide/pod-id-agent-setup.html and see if it applies to your use case?