Is the ResourceQuota admission plugin enabled by default in k3s install?

Trying to launch K3's with Ansible and I am getting Python errors. Does anyone have any ideas what is going on? Python is installed on all machines.

Any idea how to fix

unable to select an IP from default routes

error preventing k3s startup?

I’m trying to get k3s use image registry mirror, but apparently the images either are not pulled from correct registry (maybe issue with the rewrite..) or they are pulled from dockerhub. `crictl info`:

"registry": {
      "configPath": "",
      "mirrors": {
        "*": {
          "endpoint": [
            "<https://myregistry.net>"
          ],
          "rewrite": {
            "^/(.*)": "some-namespace/$1"
          }
        },
        "<http://docker.io|docker.io>": {
          "endpoint": [
            "<https://myregistry.net>"
          ],
          "rewrite": {
            "^/(.*)": "some-namespace/$1"
          }
        }
      }

I would like to manage all images from internal registry, including kube-system images.

Hi, I have a k3s cluster with 3 master nodes (embedded etcd). I converted once from a single sqlite master node. Is there a way to get back to the single master with sqlite as storage without starting a new cluster from scratch?

Hey I have a strange scenario where a cluster deployed at a site is going to be moved to another site with a different IP schema. The problem is that we do not manage the network at this site. ideally I would send someone to just rebuild the cluster but I am trying to avoid doing that. I know that this can be done by stopping kubernetes on the master node, making some config changes (deleting some stuff) and bringing it back up. Then you are able to rejoin your nodes to the master this should bring everything back. The problem is that I am currently running K3os and with it being immutable im not sure if this is even a valid option. Is there a way to just re-start the bootstrap part at the startup of K3os and I can just serve the files over the internet?

Hey y’all hopefully quick question. I’m trying to automate the installation of a k3s Rancher Server cluster with Ansible and have questions about the embedded etcd init functionality (thread)

Is it possible to override the ContainerdConfigTemplate? I’d like to provide own config file instead of template, and also use version 2 syntax (currently getting warning in logs because k3s generates config file with old syntax). Ref: https://rancher.com/docs/k3s/latest/en/advanced/#configuring-containerd

How do I enable mount propagation on my alpine linux k3s installation to make it work with longhorn storage? I tried to deploy longhorn but the manager container is not starting up due to this mount propagation error. Any help would be appreciated. @high-waitress-66594 I think perhaps you can provide some insights on how to solve this.

I am having an issue getting k3s to delete everything, which is strange. Background: I've used the official install shell script to install directly onto an ubuntu workstation, and am using the following commands to wipe the machine

sudo systemctl stop k3s
sudo k3s-killall.sh
docker rm -f $(docker ps -aq)
sudo rm -rf /var/lib/rancher /var/lib/kubelet

However, when I start k3s again, there are containers from a helm chart I was playing with that are still being created. What is the proper way to wipe everything?

I’m trying to update a cluster to v1.23.7+k3s1 from v.1.23.6+k3s1. However, when the first control-plane node comes up it can’t start and

journalct -u k3s.service -f

shows:

Jul 01 20:25:25 test-test-master-0 systemd[1]: Failed to start Lightweight Kubernetes.
Jul 01 20:25:30 test-test-master-0 systemd[1]: k3s.service: Scheduled restart job, restart counter is at 151.
Jul 01 20:25:30 test-test-master-0 systemd[1]: Stopped Lightweight Kubernetes.
Jul 01 20:25:30 test-test-master-0 systemd[1]: Starting Lightweight Kubernetes...
Jul 01 20:25:30 test-test-master-0 sh[53137]: + /usr/bin/systemctl is-enabled --quiet nm-cloud-setup.service
Jul 01 20:25:30 test-test-master-0 sh[53138]: Failed to get unit file state for nm-cloud-setup.service: No such file or directory
Jul 01 20:25:30 test-test-master-0 k3s[53141]: time="2022-07-01T20:25:30+02:00" level=info msg="Starting k3s v1.23.7+k3s1 (ec61c667)"
Jul 01 20:25:30 test-test-master-0 k3s[53141]: time="2022-07-01T20:25:30+02:00" level=warning msg="Cluster CA certificate is not trusted by the host CA bundle, but the token does not include a CA hash. Use the full token from the server's node-token file to enable Cluster CA validation."
Jul 01 20:25:30 test-test-master-0 k3s[53141]: time="2022-07-01T20:25:30+02:00" level=info msg="Managed etcd cluster not yet initialized"
Jul 01 20:25:30 test-test-master-0 k3s[53141]: time="2022-07-01T20:25:30+02:00" level=warning msg="Cluster CA certificate is not trusted by the host CA bundle, but the token does not include a CA hash. Use the full token from the server's node-token file to enable Cluster CA validation."
Jul 01 20:25:30 test-test-master-0 k3s[53141]: time="2022-07-01T20:25:30+02:00" level=fatal msg="starting kubernetes: preparing server: failed to validate server configuration: critical configuration value mismatch"
Jul 01 20:25:30 test-test-master-0 systemd[1]: k3s.service: Main process exited, code=exited, status=1/FAILURE
Jul 01 20:25:30 test-test-master-0 systemd[1]: k3s.service: Failed with result 'exit-code'.
Jul 01 20:25:30 test-test-master-0 systemd[1]: Failed to start Lightweight Kubernetes.

Is kine being used at all when using external etcd datastore in k3s cluster? Just saw kine 0.9.3 in rc, so was wondering if I need to upgrade or it is not necessary at all, because am using external etcd datastore. Also, how to clear the external etcd datastore completely if I am re-deploying k3s on a fresh install?

Hi! I'm new to k3s, coming from Docker Swarm in a small home lab environment. In my k3s deployment, all (3) nodes have shared storage already, located under

/mnt/storage

. For this reason, I don't want to use Longhorn. I'd like to use my existing replicated storage. k3s seems to default to

/var/lib/rancher/k3s/storage

with the local storage provider. I see an option for use during initial k3s setup,

--default-local-storage-path

per the documentation at https://rancher.com/docs/k3s/latest/en/installation/install-options/server-config/#storage-class The setting appears to be located in

/var/lib/rancher/k3s/server/manifests/local-storage.yaml

on each node, but it also appears from reading around on forums that this gets overwritten and should be changed with tools, rather than directly. • Is it possible to change the

--default-local-storage-path

parameter without reinstalling k3s? • Is this even the correct parameter to change? Thanks!

I’m trying to get stargz work with private image registry when using k3s and containerd. I got the authentication working using deprecated syntax (

[plugins."io.containerd.snapshotter.v1.stargz".registry.mirrors."*"]

and

[plugins."io.containerd.grpc.v1.cri".registry.mirrors."*"]

; not sure if I need both..?). I’d like to get the authentication work using also the new

config_path

&

certs.d

syntax. Does anyone have insight regarding what’s correct and recommended way? I’ve been checking docs of stargz and containerd/cri, but am a bit lost. Also, although the services eventually boot successfully, I’m currently getting lot of errors, like these during k3s install:

time="2022-07-02T14:51:25.485441124+09:00" level=info msg="Received status code: 401 Unauthorized. Refreshing creds..." key="<http://k8s.io/25/extract-737524037-rTpL|k8s.io/25/extract-737524037-rTpL> sha256:47539da01eebacb627943ac3a63d918b69534684b239edac42ce4c64742fb4fd" mountpoint=/var/lib/rancher/k3s/agent/containerd/io.containerd.snapshotter.v1.stargz/snapshotter/snapshots/17/fs parent="<http://k8s.io/24/sha256:114c5ec8519affa4cba972ba546f900b4803d5863f5242a2daee68aa6133fd7d|k8s.io/24/sha256:114c5ec8519affa4cba972ba546f900b4803d5863f5242a2daee68aa6133fd7d>" src="<http://docker.io/rancher/klipper-helm:v0.7.3-build20220613/sha256:5b8b2ba8c050ccd08bd8f00f15210c40a3f50674a0b9457b3f06870723889cd0|docker.io/rancher/klipper-helm:v0.7.3-build20220613/sha256:5b8b2ba8c050ccd08bd8f00f15210c40a3f50674a0b9457b3f06870723889cd0>"

time="2022-07-02T14:51:26.853585280+09:00" level=warning msg="failed to prepare remote snapshot" error="failed to resolve layer: failed to resolve layer \"sha256:5b8b2ba8c050ccd08bd8f00f15210c40a3f50674a0b9457b3f06870723889cd0\" from \"<http://docker.io/rancher/klipper-helm:v0.7.3-build20220613\|docker.io/rancher/klipper-helm:v0.7.3-build20220613\>": failed to resolve the blob: failed to resolve the source: cannot resolve layer: failed to redirect (host \"<http://registry-1.docker.io|registry-1.docker.io>\", ref:\"<http://docker.io/rancher/klipper-helm:v0.7.3-build20220613\|docker.io/rancher/klipper-helm:v0.7.3-build20220613\>", digest:\"sha256:5b8b2ba8c050ccd08bd8f00f15210c40a3f50674a0b9457b3f06870723889cd0\"): failed to access to the registry with code 404: failed to redirect (host \"<http://registry-jpe2.r-local.net|registry-jpe2.r-local.net>\", ref:\"<http://docker.io/rancher/klipper-helm:v0.7.3-build20220613\|docker.io/rancher/klipper-helm:v0.7.3-build20220613\>", digest:\"sha256:5b8b2ba8c050ccd08bd8f00f15210c40a3f50674a0b9457b3f06870723889cd0\"): failed to access to the registry with code 401: failed to resolve: failed to resolve target" key="<http://k8s.io/25/extract-737524037-rTpL|k8s.io/25/extract-737524037-rTpL> sha256:47539da01eebacb627943ac3a63d918b69534684b239edac42ce4c64742fb4fd" parent="<http://k8s.io/24/sha256:114c5ec8519affa4cba972ba546f900b4803d5863f5242a2daee68aa6133fd7d|k8s.io/24/sha256:114c5ec8519affa4cba972ba546f900b4803d5863f5242a2daee68aa6133fd7d>" remote-snapshot-prepared=false

time="2022-07-02T14:51:37.262725372+09:00" level=error msg="ContainerStatus for \"1e7c670c740754bfa1d6e3c4ca2540025adb6cd8063abeb057911f9535d0f1b3\" failed" error="rpc error: code = NotFound desc = an error occurred when try to find container \"1e7c670c740754bfa1d6e3c4ca2540025adb6cd8063abeb057911f9535d0f1b3\": not found"
time="2022-07-02T14:51:37.263923059+09:00" level=error msg="ContainerStatus for \"5af6495a6eeb8d647ce97e2ece46489961799cb711a85533fd4a38892b0af32e\" failed" error="rpc error: code = NotFound desc = an error occurred when try to find container \"5af6495a6eeb8d647ce97e2ece46489961799cb711a85533fd4a38892b0af32e\": not found"

Is this normal with private registries, or do I possibly have some misconfiguration somewhere…?

Is it possible to specify helm chart versions for k3s?

Hi all. I have been using an nginx ingress controller with

hostnetwork: true

and a public IP and it has been doing exactly what I want but now I'd like to put it behind an

sslh

so I can use one of the ports for ssh also. I can't seem to find how to set the bind address to

127.0.0.1

or the private IP so

sslh

can listen on the external IP and proxy to the ingress. Does anyone know what I need to set?

Here’s a funny one… I have two master nodes that disagree about which one is ready.

root@dp6448:~# k3s kubectl get nodes
NAME     STATUS     ROLES                       AGE   VERSION
dp6448   Ready      control-plane,etcd,master   69d   v1.23.4+k3s1
dp6449   NotReady   control-plane,etcd,master   69d   v1.23.4+k3s1

root@dp6449:~# k3s kubectl get nodes
NAME     STATUS     ROLES                       AGE   VERSION
dp6448   NotReady   control-plane,etcd,master   69d   v1.23.4+k3s1
dp6449   Ready      control-plane,etcd,master   69d   v1.23.4+k3s1

If I ask them to

kubectl describe node

each other, I get

Conditions:
  Type             Status    LastHeartbeatTime                 LastTransitionTime                Reason              Message
  ----             ------    -----------------                 ------------------                ------              -------
  MemoryPressure   Unknown   Tue, 05 Jul 2022 10:32:22 +0000   Tue, 05 Jul 2022 10:33:42 +0000   NodeStatusUnknown   Kubelet stopped posting node status.
  DiskPressure     Unknown   Tue, 05 Jul 2022 10:32:22 +0000   Tue, 05 Jul 2022 10:33:42 +0000   NodeStatusUnknown   Kubelet stopped posting node status.
  PIDPressure      Unknown   Tue, 05 Jul 2022 10:32:22 +0000   Tue, 05 Jul 2022 10:33:42 +0000   NodeStatusUnknown   Kubelet stopped posting node status.
  Ready            Unknown   Tue, 05 Jul 2022 10:32:22 +0000   Tue, 05 Jul 2022 10:33:42 +0000   NodeStatusUnknown   Kubelet stopped posting node status.

Everything other than that seems fine… etcd is healthy (one of the nodes is the leader and the other is not), they’re just both surfacing that this node isn’t ready. Confused as to how this is possible with healthy etcd.

curl <https://get.k3s.io/> | INSTALL_K3S_VERSION="v1.23.6+k3s1" INSTALL_K3S_EXEC="--disable=traefik --node-ip 10.2.0.1" sh -

curl <http://10.42.0.4:9153>

works (gives 404, ok). But on a neighbor host (10.2.0.210),

curl <http://10.2.0.1:9153>

is connection-refused. Shouldn't that work? where do i debug next?

For the etcd backups, I should run all 3 k3s instances with the same flags right? Deploying to just one doesn’t setup etcd backups on all of them? Also, do they coordinate or are they all 3 competing to put backup their own version of the world at the same time? Do they just let the leader do it?

Hello - Is there a K3s equivalent for a path location where certs are stored? E.g. /etc/kubernetes/pki/ ?

Where can I find HelmChart CRD definition (preferably Git)?

fyi, it seems the info related to kubernetes-dashboard does not apply anymore. Also apparently need kubectl 1.24+ in order to create the token.

Hi all, I have a k3s cluster with 2 nodes(1GB RAM and 2 CPUs). When I run some compute intensive workload on the cluster the node goes to a not ready state(because of OOM). I want to avoid the node going to not ready state instead I want to kill the pod which is consuming more resources. I tried setting the

kube-reserved, system-reserved, kube-reserved-cgroup, system-reserved-cgroup

flags on kubelet. But If I don't have requests or limits in my pod spec it still crashes the node(NotReady state). I have tried combination of setting the above flags with eviction-hard flag as well. Is there any other solution to overcome this?

Opened PR for documentation https://github.com/rancher/docs/pull/4185

Hi all, guess I'm pretty late checking out k3s. It's wonderfully lightweight. I'm trying to

go get <http://github.com/k3s-io/k3s|github.com/k3s-io/k3s>

but getting error

go: downloading <http://github.com/k3s-io/k3s|github.com/k3s-io/k3s> v1.21.9
go get: <http://github.com/k3s-io/k3s@v1.21.9|github.com/k3s-io/k3s@v1.21.9>: parsing go.mod:
	module declares its path as: <http://github.com/rancher/k3s|github.com/rancher/k3s>
	        but was required as: <http://github.com/k3s-io/k3s|github.com/k3s-io/k3s>

I then tried

go get <http://github.com/rancher/k3s|github.com/rancher/k3s>

but again it errors

go: downloading <http://github.com/rancher/k3s|github.com/rancher/k3s> v1.21.9
go get: <http://github.com/rancher/k3s@v1.21.9|github.com/rancher/k3s@v1.21.9> requires
	<http://github.com/kubernetes-sigs/cri-tools@v0.0.0-00010101000000-000000000000|github.com/kubernetes-sigs/cri-tools@v0.0.0-00010101000000-000000000000>: invalid version: unknown revision 000000000000

How can I

go get

like most go modules?

IIRC, it was possible to write a

<http://upgrade.cattle.io/v1|upgrade.cattle.io/v1>

plan, for Ubuntu packages upgrades. I think at least I once saw an example of that, but my search fu is lacking. Any known pointers to such a thing?

Hi Team, I have upgraded

k3s

version from

1.20.7+k3s1

to

1.22.8+k3s1

I see some changes like image name being named as

rancher/mirrored-coredns-coredns:1.9.1

for COREDNS pod and

rancher/mirrored-library-traefik:2.6.1

for TRAEFIK pod. Is this an expected or am i missing some configuration or image being used is incorrect ? Please note both are part

k3s

installation and not installed separately... TIA

Does k3s (including all the components, e.g. containerd) run on btrfs filesystem? If so, are there benefits or possible issues in such compared to ext4?

Hi folks. I’m comparing a few different Kubernetes variants, looking into how much memory they use per CRD installed. Note that I mean literally per CustomResourceDefinition that is created - not per custom resource. I’m seeing

k3s server

w/sqlite use ~4MB/CRD, which is about the same as the upstream API server. Does that sound about right?