rancher-users #k3d

adamant-ram-8166

05/02/2022, 7:22 PM

Feeling really dumb here. Created a cluster with the following commands on my Macbook Pro M1 and am able to connect over localhost just fine, but when I try it on my Ubuntu desktop I can't connect. I've got Rancher Desktop installed and configured with docker/moby on both. 🧵

bulky-appointment-8113

05/04/2022, 10:56 AM

Hey, I want to use k3d in a Github action for e2e tests. Anyone have a preference on https://github.com/AbsaOSS/k3d-action vs https://github.com/nolar/setup-k3d-k3s ? AbsaOSS looks a bit more recent?

orange-ocean-99788

05/05/2022, 12:42 AM

I am deploying a cluster with 3 node apple M1 macs, and noticed that “distributed cluster” was moved to milestone 5.5, hoping for this enhancement asap.

millions-alarm-86298

05/06/2022, 9:03 AM

skaffold + k3d is unbeatable combo for CI / CD and e2e testing of helm charts very smoothly running on GH Actions and allows doing "crazy things" 😉

🙏 1

🎉 1

glamorous-engine-17369

05/07/2022, 1:54 PM

So I'm trying to deploy cert-manager on a bare k3d server, and I'm getting issues with cert-manager's ca-injector not being able to perform leader election, is this a known issue with k3d? I think I had it working before

millions-alarm-86298

05/09/2022, 7:48 PM

Hi all I see this issue is closed https://github.com/k3d-io/k3d/issues/122 But I can't find related topic in docs https://k3d.io/v5.0.0/ Am I missing smth ?

millions-alarm-86298

05/10/2022, 5:39 PM

Hi all I never used internal k3d traefik before, so encountered an issue when creating a new cluster with LB and traefik the

k3d cluster create

doesn't wait until traefik is up , so I manually must check if helm chart and CRDs are installed is there any way to improve this experience ?

brief-ability-87157

05/11/2022, 12:55 AM

Hello everyone! How can I add a private registry to an existing cluster? From the doc on the site, it's not clear if it's possible. Initially, I created a cluster with

--registry-config

flag.

brief-ability-87157

05/11/2022, 1:26 AM

I added a register in the docker container

{name}-server-0

directly but wondering if there is another way.

flat-florist-77376

05/11/2022, 3:58 PM

Hey all--anyone know why when I create a cluster with one node with the

--host-alias

flag, that works… unless I restart the node, in which case the alias I configured earlier goes missing?

glamorous-elephant-6949

05/14/2022, 5:47 AM

I've got an issue with k3d on ArchLinux with my firewalld configuratuion.

kubectl cluster-info

returned a valid config but

kubectl get pods

showed all bootstrap containers in

CreatingContainer

state.

alert-oxygen-95787

05/16/2022, 2:01 PM

Is it better to use k3s with multipass or docker (via k3d) on a Mac M1? Using multipass VMs seems a bit closer to how k8s is deployed in production as at least you can interact with the individual nodes. Just curious if anyone has noticed any advantages/disadvantages with either approach. Thanks!

wooden-coat-97755

05/17/2022, 2:00 PM

👋🏻 k3d experts! I’m looking into k3d and found it interesting that it bundles an nginx container in front of k3s clusters by default. Can anybody explain the reasoning behind that design decision?

green-musician-2686

05/17/2022, 9:10 PM

Hi all, does anybody use k3d in LCX/LXD containers ?

red-afternoon-63849

05/19/2022, 4:49 PM

Hi there, I'm working with

k3d

and tilt.dev on a helmchart that requires

sysbox

and

CRI-O

is there anyway to get k3d to utilize a

CRI-O

runtime?

handsome-answer-69741

05/20/2022, 7:36 AM

Good day, I am manually creating k3d cluster on a fresh AWS EC2 instance. 1. Installed docker, k3s, k3d. 2. Run k3d command to create cluster with 3 + 3 nodes (k3d cluster create my-cluster --servers 3 --agents 3). Issue- when performing kubectl get nodes, I assume to list all created k3d nodes, instead, I get EC2 instance. When running k3d node ls, I clearly see all created nodes. Question- how can I work in the k3d cluster, example, install helm charts in cluster instead of on EC2 instance. Many thanks!

melodic-market-42092

06/07/2022, 11:22 AM

Are network policies applied/in effect by default with k3d when I start the cluster with

k3d cluster create $CLUSTER_NAME

and without any configuration files?

bumpy-laptop-27086

06/15/2022, 10:24 AM

Hi folks. My team is intending to adopt k3d for a local development environment, and we have to deploy infrastructure tools such as PostgreSQL, Kafka, Prometheus and so on. I’ve already been able to deploy all of them, but I’m struggling to enable access from the host machine (esp. kafka). I’m wondering if somebody here did it before and if so, could they point me towards the right path? Normally what I would do is to setup a load balancer, but this doesn’t work, I suspect is because of Docker for Mac. Thanks in advanced and have a nice day.

✅ 1

wooden-potato-2675

06/16/2022, 2:22 PM

Copy code

FATA[0018] Failed to start server k3d-poccluster-server-2: Node k3d-poccluster-server-2 failed to get ready: error waiting for log line `k3s is up and running` from node 'k3d-poccluster-server-2': stopped returning log lines

proud-eve-37927

06/20/2022, 8:48 PM

Hi all, I try to switch from Kind to k3d. But I fail in the very beginning because I am using flux to deploy my stuff. If I deploy the following manifest:

apiVersion: <http://source.toolkit.fluxcd.io/v1beta1|source.toolkit.fluxcd.io/v1beta1>

kind: HelmRepository metadata: name: vault spec: interval: 5m url: https://helm.releases.hashicorp.com I got the following error from flux:

❯ kg <http://helmrepositories.source.toolkit.fluxcd.io|helmrepositories.source.toolkit.fluxcd.io> vault

NAME URL AGE READY STATUS vault https://helm.releases.hashicorp.com 10m False failed to fetch Helm repository index: failed to cache index to temporary file: Get "https://helm.releases.hashicorp.com/index.yaml": dial tcp: lookup helm.releases.hashicorp.com on 10.43.0.1053 server misbehaving on the Kind / rke / minikube it works perfect. I have no glue what's going wrong there in k3d. Can anybody help me please in this issue Thanks a lot Klaus

✅ 1

melodic-market-42092

06/23/2022, 1:49 PM

I'm using https://k3d.io/v5.4.1/usage/exposing_services/ and

k3d cluster create $CLUSTER_NAME -p "8081:3000@loadbalancer"

to create my local cluster and get something exposed. This works well, but now I'd like to expose two things. Is that possible? Can I do something like

k3d cluster create $CLUSTER_NAME -p "8081:3000,8082:8000@loadbalancer"

melodic-market-42092

06/24/2022, 10:00 AM

I'm doing

Copy code

k3d cluster create $CLUSTER_NAME \
    -p "8081:3000@loadbalancer" \
    -p "8082:8000@loadbalancer" \
    -p "8086:8086@loadbalancer" \
    -p "8087:2746@loadbalancer"

to expose a few things from my k3d cluster. The first three of those work, but the last one just.. doesn't. I'm struggling to figure out why. If I do a port-forward to the service, then things work as expected. The ingress and service for the fourth entry there (8087:2746) looks exactly similar to the ingress'es for the other three 🤔 The application behind 2746 is https://argoproj.github.io/argo-workflows/argo-server/. Any hints anybody could give on this? Hopefully someone has seen this before.

cool-sunset-23736

07/17/2022, 4:46 PM

What I don't get is how the configs get written into the

/etc/confd/values.yaml

file. Is this a shared volume between the nodes or is there some other mechanism in play?

millions-alarm-86298

07/20/2022, 7:10 AM

Hello Looks like I found a regression in 5.4.4 1. create k3d cluster 2. deploy smth there - not sure what because I deploy few tools includinig custom traefik 3. restart PC / laptop 4. docker container from this image rancher/k3s:v1.23.6-k3s1 is constantly failing to start 5. this causes kubectl to work unstable, with lags etc ( apparetnly because API is unavailable ) 5.4.3 works fine

fierce-room-79104

07/24/2022, 12:48 PM

Hiya, question; Is there anyone over here who's got k3d to work with rootless podman on archlinux?

fierce-room-79104

07/24/2022, 12:49 PM

I'm currently failing to create a cluster:

Copy code

ERRO[0005] Failed Cluster Start: Failed to start server k3d-foo-server-0: Node k3d-foo-server-0 failed to get ready: Failed waiting for log message 'k3s is up and running' from node 'k3d-foo-server-0': node 'k3d-foo-server-0' (container 'd0fbe81acb8e3fb4c0e7e051c65b93346cb6180d88ec8cd84301a9282c71e7c6') not runningERRO[0005] Failed to create cluster >>> Rolling BackINFO[0005] Deleting cluster 'foo'DEBU[0005] Cluster Details: &{Name:foo Network:{Name:k3d-foo ID:b7842ddccf7c6454bd9297f3b4c977cbf1cae3de909c6fef95af3aa5c10e9334 External:false IPAM:{IPPrefix:10.89.0.0/24 IPsUsed:[] Managed:false} Members:[]} Token:SDuOUscMoGAPLaLcgqkR Nodes:[0xc00008b6c0 0xc00008b860] InitNode:<nil> ExternalDatastore:<nil> KubeAPI:0xc000041080 ServerLoadBalancer:0xc0001ffd30 ImageVolume:k3d-foo-images Volumes:[k3d-foo-images k3d-foo-images]}DEBU[0005] Deleting node k3d-foo-serverlb ...DEBU[0005] Deleting node k3d-foo-server-0 ...INFO[0005] Deleting cluster network 'k3d-foo'INFO[0005] Deleting 2 attached volumes...DEBU[0005] Deleting volume k3d-foo-images...DEBU[0005] Deleting volume k3d-foo-images...WARN[0005] Failed to delete volume 'k3d-foo-images' of cluster 'foo': failed to find volume 'k3d-foo-images': Error: No such volume: k3d-foo-images -> Try to delete it manuallyFATA[0005] Cluster creation FAILED, all changes have been rolled back!

And on the podman side it logs:

Copy code

WARN[0016] Could not find mount at destination "/var/run" when parsing user volumes for container 498e3a3624702a4de4b372373774cffa26d8b77390aafbaceb4b7a18e596dfddWARN[0017] Could not find mount at destination "/var/run" when parsing user volumes for container 498e3a3624702a4de4b372373774cffa26d8b77390aafbaceb4b7a18e596dfddWARN[0017] Could not find mount at destination "/var/run" when parsing user volumes for container 498e3a3624702a4de4b372373774cffa26d8b77390aafbaceb4b7a18e596dfddWARN[0017] Could not find mount at destination "/var/run" when parsing user volumes for container 498e3a3624702a4de4b372373774cffa26d8b77390aafbaceb4b7a18e596dfddWARN[0017] Could not find mount at destination "/var/run" when parsing user volumes for container 498e3a3624702a4de4b372373774cffa26d8b77390aafbaceb4b7a18e596dfdd

big-librarian-12753

08/01/2022, 10:14 AM

hi, I think it may have come up many times, still trying one more time. Is it possible anyway to make k3d use local docker cache via any proxy or anything else which makes my life easier rather than importing images everytime into the cluster?

millions-alarm-86298

08/03/2022, 10:55 AM

k3d is mentioned in a 1st place https://traefik.io/blog/top-5-open-source-tools-local-development-kubernetes/?utm_[…]&utm_medium=social&utm_source=twitter&hss_channel=tw-4890312130

melodic-market-42092

08/03/2022, 12:16 PM

If I modify my

/etc/hosts

like this:

Copy code

localhost       myapp.local

And have an ingress like this:

Copy code

spec:
  rules:
    - host: myapp.local

And run k3d like this:

Copy code

k3d cluster create $CLUSTER_NAME \
    -p "80:80@loadbalancer" \

Should I then be able to reach the service that my ingress points to by going to http://myapp.local in the browser?

many-telephone-82541

08/09/2022, 6:11 PM

I have issue with testing the Vault CSI Secrets driver on k3d on Mac. I get the following message:

Error: failed to generate container "69815de67a72a047842260d67a1d83fa61f6cf073a3cbf4b09fc1166926cec26 │

│ " spec: failed to generate spec: path "/var/lib/kubelet/pods" is mounted on "/var/lib/kubelet" but it is not a shared mount

Any suggestions?