I am facing the same issue here, the logs says that the dns is not resolved and the connection times out.

2024/12/19 10:00:42 [error] 17#17: recv() failed (111: Connection refused) while resolving, resolver: 127.0.0.11:53
2024/12/19 10:00:47 [error] 17#17: recv() failed (111: Connection refused) while resolving, resolver: 127.0.0.11:53
2024/12/19 10:00:52 [error] 17#17: recv() failed (111: Connection refused) while resolving, resolver: 127.0.0.11:53
2024/12/19 10:00:57 [error] 17#17: recv() failed (111: Connection refused) while resolving, resolver: 127.0.0.11:53
2024/12/19 10:01:02 [error] 17#17: recv() failed (111: Connection refused) while resolving, resolver: 127.0.0.11:53
2024/12/19 10:01:07 [error] 17#17: recv() failed (111: Connection refused) while resolving, resolver: 127.0.0.11:53
2024/12/19 10:01:12 [error] 17#17: *47 minio.REDACTED could not be resolved (110: Operation timed out) while sending to client, client: 10.244.130.151, server: _, request: "GET /assets/by-id/1d7970d0-f032-45e4-9227-f199eda296d7 HTTP/1.1", host: "REDACTED", referrer: "https://REDACTED/"
10.244.130.151 - - [19/Dec/2024:10:01:12 +0000] "GET /assets/by-id/1d7970d0-f032-45e4-9227-f199eda296d7 HTTP/1.1" 502 552 "https://REDACTED/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36"

I have prepared a PR to resolve this. This can be deployed in next version.

However if you need you can use the fork that i created and use the devel readme to deploy the modified chart for now.

Thank you for your contribution, @vipinjn24 and @a3nick. It will be available with the next release (hopefully tomorrow). But I think this solution does not work for some scenarios, for example using replicas, and It would be great if a suitable value could be set ‘automatically’ for each pot. We will work on it, but we are open to any suggestions.

I think I found a solution to set the default IP of the Pod so that you don't have to set it by hand. Try it out and let me know if it works for you. In my environment I can't reproduce the bug.

I will close the issue but feel free to reopen it if the error persists.

I think I found a solution to set the default IP of the Pod so that you don't have to set it by hand. Try it out and let me know if it works for you. In my environment I can't reproduce the bug.

I will close the issue but feel free to reopen it if the error persists.

I tried the new helm chart version.

Using the config.internalResolver value works fine by specifying the ip address of the main k8s nameserver.
Usually, for CoreDns, is x.x.x.10, where x.x.x. is the base internal k8s CIDR

The solution you proposed, using the pod internal ip, doesn't work

2025/01/09 12:58:42 [error] 9#9: recv() failed (111: Connection refused) while resolving, resolver: 10.42.0.88:53

nothing is running on port 53, so the call fails