Heyo, i have a little trouble with my k3d pods reaching a certain website. I run the cluster with

--network host

and checked that the resolv.conf has the correct search domains. The host machine can reach the domain. The pod cannot and but if it can ping the ip

Can someone help me understand why that can happen and or what the network host parameter does in detail?

whats the best way to know the latest supported k3s version with k3d ?

I am confused with with

+

and

-

variants ;D

It should work with any version

Those aren't variants. You just can't have a + in image tags, Docker doesn't allow it. Any place there's a + in the version, use a - in the tag.

I have created a cluster with

apiVersion: <http://k3d.io/v1alpha4|k3d.io/v1alpha4>
kind: Simple
metadata:
  name: gitops-quickstart
servers: 1
agents: 1
image: rancher/k3s:v1.24.4-k3s1
network: gitops-quickstart
ports:
  # Drone CI
  - port: 0.0.0.0:30980:30980
    nodeFilters:
      - loadbalancer
  # Gitea
  - port: 0.0.0.0:30950:30950
    nodeFilters:
      - loadbalancer
  # Argo CD
  - port: 0.0.0.0:30080:30080
    nodeFilters:
     - loadbalancer

Hi friends! When creating a new registry, I always get the prefix

k3d

. Can I override that somehow? When I create a cluster and registry in one go, I can set any name:

k3d cluster create --registry-create foo:127.0.0.1:1234

But when creating the registry alone, I always get the

k3d

prefix:

k3d registry create foo --port 127.0.0.1:1234

@bland-area-30120 has left the channel

Hello folks. I did quick search for this but no luck. Is there a way to get k3d to use docker’s credential helper when trying to pull the k3s image? I’m using a custom k3s image in GKE’s Artifact Store and it fails with the message

Permission "artifactregistry.repositories.downloadArtifacts" denied on resource

. A manual docker pull works fine so I’m assuming k3d can’t use the gcloud credential helper.

Howdy, I'm trying to modify coredns to forward upstream request to a specific server, but that setting doesn't seem to stick reliably. I'm looking at the modifications using a configmap called coredns-custom. I see a few issues regarding this, but I've not seen documentation. Is there something more needed than creating the configmap? Perhaps mounting it within the container?

@wide-garage-9465 do we have examples on how to use

rancher/k3d:x-dind

image ? I want to start a k3d cluster without mounting host docker sock

why there is no checksum for https://github.com/k3d-io/k3d/releases ?

any thoughts on https://rancher-users.slack.com/archives/CGGQEHPPW/p1675591781414439?

We're using k3d locally and are upgrading from 4.4.6 to 5.4.6. With 4.4.6 we mounted an updated coredns configmap to /var/lib/rancher/k3s/server/manifests/d-coredns-patch.yaml so that we could add

rewrite name regex (.*).<http://local.example.com|local.example.com> public-nginx-ingress-nginx-controller.default.svc.cluster.local

. (Everything else was kept default.) It appears from my reading of github issues that modifying the

.:53

block in coredns is no longer supported? Any pointers on how to handle this? For now I'm doing this, but it feels... wrong.

kubectl -n kube-system patch configmap coredns --patch-file "coredns-patch.yaml"
kubectl wait --for=condition=Ready=true pod -l k8s-app=kube-dns -n kube-system
kubectl -n kube-system rollout restart deployment coredns

Is there a list of storage providers known to work with k3d/k3d?

hi! I try to understand the k3d (k3s). It is a lightweight kubernetes, it is okay. https://k3d.io/ Guides So, On a machine (laptop) I can run it with docker. Pre-requirements: docker, kubectl The "install.sh" setup the environment: cluster with nodes. Master Node (Control Plane): Loadbalancer + Nginx Proxy Worker Nodes: k3d-test-server-0 ... k3d-test-agent-0 ... what are the k3d-test-server-X and k3d-test-agent-X? What is the difference the server and agent?

K3D has no web interface? Can I use only the K3X?

Good day. When I install k3d, create a default cluster and then run

kubectl cluster-info

I get a couple of messages like

memcache.go:255] couldn't get resource list for <http://metrics.k8s.io/v1beta1|metrics.k8s.io/v1beta1>: the server is currently unable to handle the request

I have (what I hope) is a really easy question. I'm launching a k3d cluster using the following command:

k3d cluster create doctorconsul --network doctorconsul_wan --api-port 127.0.0.1:6443 --k3s-arg="--disable=traefik@server:0" -p "8502:443@loadbalancer"

This works pretty much exactly as expected. It drops into my "doctorconsul_wan" network that is already built in docker compose. The problem I'm having is that the external-IP used is dynamically the first available in that network (192.168.1.7 in my case) and I need to statically assign it (to 192.169.1.10). I'm not seeing a parameter to do so. I have a feeling that perhaps I have to pass a kube native config param through via

--kubeconfig-update-default

, but frankly I don't know nearly enough about kube to figure that out. Any help is much appreciated. Thanks!

Alright. So modifying the server node IPs appears to have NOT actually worked. it changes the server IPs, but everything is broken. when I use this syntax:

k3d cluster create doctorconsul --network doctorconsul_wan \
    --api-port 127.0.0.1:6443 \
    -p "8502:443@loadbalancer" \
    --k3s-arg="--disable=traefik@server:0" \
    --k3s-arg="--node-ip=192.169.7.10@server:0" \
    --k3s-arg="-node-external-ip=192.169.7.10@server:0"

The only server node does correctly have these IPs

Addresses:                                                                                                                                                                                                                                                                ││   InternalIP:  192.169.7.10                                                                                                                                                                                                                                               ││   ExternalIP:  192.169.7.10                                                                                                                                                                                                                                               ││   Hostname:    k3d-doctorconsul-server-0

But coreDNS won't come up.

[INFO] plugin/ready: Still waiting on: "kubernetes"                                                                                                      
[WARNING] plugin/kubernetes: Kubernetes API connection failure: Get "<https://10.43.0.1:443/version>": dial tcp 10.43.0.1:443: connect: no route to host

^^^ These are just repeatedly spammed. Anyone know why? I'm way out of my kube league and I'm just trying to get K3d to accept a static IP instead of choosing the first avail dynamically. Thanks

Hi Everyone, hope you guys are doing great. This is my first time installing k3d and what I'm seeing is that even though k3d installation was successful, all the pods are in pending state. Am I missing something here?

I'm trying to use k3d to run an application on some shared lab (read "wild west with nearly everyone having root") machines. The owners of the machines are required to do some vulnerability scans that are finding things inside /var/lib/docker/volumes , overlay2 , etc and are sending out "hey guys, fix these" e-mails. they took a pass at trying to translate from the location to the image container and some of them were definite offenders but others were unrelated. I think the method they used was just with the docker command. I just spent an hour or two poking around trying to get at the internals of which container were what using docker & ctr commands with a k3d cluster running and wasn't having a lot of luck. Anyone know if I could've gotten to things if I'd just used nerdctl or crictl or checked the right spot in kubectl or some other tool? I didn't spot anything on k3d.io and thought I'd ask if anyone knows before I go experimenting with the different tools or even spelunking through odd spots in Linux.

I know with RKE2 sometimes you can query some items with crictl (I think it's the static pods only). Does K3D run anything with crio such that it could be queried with crictl too?

Hello, Gm How can I enable CpuManager please on k3d ?

--k3s-server-arg '--kube-apiserver-arg=feature-gates=CPUManager=true --cpu-manager-policy=static'

maybe ?

Hi, I am trying to run a k3d cluster for the first time, and like other users, all the pods are in pending state. However, this doesn't seem to be the same problem as https://github.com/k3d-io/k3d/issues/1229, I don't see the same error. Any hints on how to debug this ? I'm on Ubuntu 22.10 with a ZFS filesystem, and after reading k3d and k3s issues with ZFS, I'm not sure whether ZFS is supported now or not ?

if I wanted to create (or edit) a k3d cluster's kube-control-manager so that I can set

--horizontal-pod-autoscaler-sync-period

, what's the proper way to do this?

Hi experts, I created a kubernetes cluster using k3d on one VM and I want to add other 20 VMs as worker nodes to this cluster. Is this feasible?