I’ll go there, i don’t constantly post questions but I’ve just been recently having alot of issues with CNI’s, I might just delete this post

I increase envoys memory requirement to 1gb, did not fix the issue.

Well I switched to cilium, same issue, and the reason I started using a CNI earlier than I intended was because flannel didn’t work.

This issue might seem complex but could you tell me some debugging stuff and logs to try to maybe get to the source of the issue or atleast provide a way to reproduce my issue (so I could maybe file a bug report)

Not k3d, just plain k3s

spiderunderurbed@raspberrypi:~/k8s $ kubectl get networkpolicy -A
No resources found
spiderunderurbed@raspberrypi:~/k8s $ 

No networkpolicies.

spiderunderurbed@raspberrypi:~/k8s $ kubectl get pods -A | grep -i dns
default                      pdns-admin-mysql-854c4f79d9-wsclq                         1/1     Running            1 (2d22h ago)    4d9h
default                      pdns-mysql-master-6cddc8cd54-cgbs9                        1/1     Running            0                7h49m
kube-system                  coredns-ff8999cc5-hchq6                                   1/1     Running            1 (2d22h ago)    4d11h
kube-system                  svclb-pdns-mysql-master-1993c118-8xqzh                    3/3     Running            0                4d
kube-system                  svclb-pdns-mysql-master-1993c118-whf5g                    3/3     Running            0                124m
spiderunderurbed@raspberrypi:~/k8s $ 

Ignore powerdns, its just extra stuff, but yeah coredns is running

spiderunderurbed@raspberrypi:~/k8s $  kubectl get endpoints  -n kube-system
NAME             ENDPOINTS                                              AGE
kube-dns         172.16.246.61:53,172.16.246.61:53,172.16.246.61:9153   4d11h
metrics-server   172.16.246.45:10250                                    4d11h
traefik          <none>                                                 130m
spiderunderurbed@raspberrypi:~/k8s $ 

^ endpoints and services:

spiderunderurbed@raspberrypi:~/k8s $ kubectl get svc -n kube-system
NAME             TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)                      AGE
kube-dns         ClusterIP      10.43.0.10      <none>        53/UDP,53/TCP,9153/TCP       4d11h
metrics-server   ClusterIP      10.43.67.112    <none>        443/TCP                      4d11h
traefik          LoadBalancer   10.43.116.221   <pending>     80:31123/TCP,443:30651/TCP   131m
spiderunderurbed@raspberrypi:~/k8s $ 

It was my backend, turns out, it forwards /nextcloud onto the nextcloud service, which does not know what to do with it unless I set something like site-url to include that path. So I made a middleware to strip the prefix, but now it cannot access any of its files because it will use the wrong path.

https://pastebin.com/RhU5xtma I cant access any external address including dns servers, so, there is no firewall running on my pi (the master node), I can set the nameserver to be 1.1.1.1 in the pods config and iirc that works, but inside the pod, it doesn’t work, so how do i fix this? You probably need more information so i can share. I am running calico as my CNI

Could you explain the steps to active your setup? I’m sure I can modify it for powerdns

Well, its kube-dns, and it simply, does not work, more specifically, it cannot resolve any external domains, I think it can resolve internal domains but I doubt thats working, but mainly it cant resolve external domains. I posted about it, here: https://lemmy.zip/post/36964791

Recently, it was fixed because I found the correct endpoint, and uhh, now it stopped working, I updated the endpoint to the newer one, but it went back to the original issue detailed in that post.

No, i want to replace kube-dns and coredns, and some of my applications will resolve the ip at my dns server, then try those ips within the server, but mainly I want to replace the current dns stack due to several issues.

Know anything that can do what I want?

I assume external dns means it will add all the domains like <pod>.<namespace>.svc.cluster.local entries to that, or just <pod>

Any idea how I do that?

Thats weird, because I clearly have free space:

spiderunderurbed@raspberrypi:~ $ df -h /
Filesystem      Size  Used Avail Use% Mounted on
/dev/mmcblk0p2  235G  184G   39G  83% /
spiderunderurbed@raspberrypi:~ $ df -h .
Filesystem      Size  Used Avail Use% Mounted on
/dev/mmcblk0p2  235G  184G   39G  83% /home
spiderunderurbed@raspberrypi:~ $ 

any idea of things I can try to fix/debug this?