DNS resolver configuration in reverse proxy should respect /etc/resolv.conf for k8s etc.. container runtime.
Closed this issue · 3 comments
Describe the bug
When I deploy this image on my Kubernetes cluster, the reverse proxy fails due to a DNS lookup error.
I didn't notice this problem because I didn't test it on kubernetes when I tested 2.0.0 in #119 🙇 .
Expected behavior
The system should resolve the hostname for Kratos in the internal network, applicable not only to Docker locally but also to other container runtimes.
Nginx should prioritize using /etc/resolv.conf, and default to 127.0.0.11 only if this configuration file is unavailable.
I confirmed /etc/resolv.conf is also available in the container on Docker for Desktop.
# cat /etc/resolv.conf
nameserver 127.0.0.11
options ndots:0
Kubernetes:
- Version 1.21
kratos-admin-ui
- Version 2.0.0
Related context
It seems nginx-ingress-controller is also taking same approach.
I will make a PR if I find some spare time in the coming days.
Thanks for that contribution! I just released v2.0.1 🎉