This message was deleted.

You set your settings in the

config.yaml

at

/etc/rancher/rke2/config.yaml

. you can also set the config options for clusters via the cluster crd under the

rkeConfig

section https://docs.rke2.io/install/configuration

Thank you, can I set it of existing cluster?

Yup - update the file then restart the rke2 service in the node

so there is no

config.yaml

at

/etc/rancher/rke2

but there is

/etc/rancher/rke2/config.yaml.d/50-rancher.yaml

but it it's not yaml file and I'm not sure it is correct one. I just need to modify

eviction-hard

to modify disk pressure thresholds on the nodes

Oh yea its different when rancher provisions it - sorry forgot about that. Yes that 50-rancher.yaml is the correct file. It's just in json instead of yaml.

hm, it does not take my setting. I've added below but then then the service would not start

"kubelet": [
    "extra_args": [
      "eviction-hard: "imagefs.available<5%,nodefs.available<5%" ]]

yea that's not quite exactly the correct syntax - so how the config.yaml file is set up is its based on the flags you pass the rke2 binary when it starts - say for example you are starting a server node - these are going to be all the flags: https://docs.rke2.io/reference/server_config most important to this is probably

--kubelet-arg

so take that flag and convert it to JSON and put it in that 50-rancher.yaml file.

"kubelet-args": [
  "eviction-hard: "imagefs.available<5%,nodefs.available<5%"
]

I'll test this too - just give me a few - its possible I messed up the syntax here

tried it but service would not start

I realized that i missed a close comma after

eviction-hard

that might have something to do with it

ok, the

"

was missing, here is that worked so far:

"kubelet-args": [
    "eviction-hard": "imagefs.available<5%,nodefs.available<5%"
  ],

now need to test it

ahh dang yea the kubelet switched to a config file for it's configurations... yea let me look it up

I've made yaml with

kind: KubeletConfiguration

but when trying to apply getting

no matches for kind "KubeletConfiguration" in version "<http://kubelet.config.k8s.io/v1beta1|kubelet.config.k8s.io/v1beta1>"

yea that wont be applied to the cluster it will be referenced by kubelet. Much like setting max pods. https://github.com/rancher/rke/issues/1298#issuecomment-1482008422

I'm losing will to live 🙂 so I have my kube-config.yaml

apiVersion: <http://kubelet.config.k8s.io/v1beta1|kubelet.config.k8s.io/v1beta1>
kind: KubeletConfiguration
evictionHard:
  nodefs.available: "1Gi"
  imagefs.available: "100Gi"
  nodefs.inodesFree: "1%"

I made a change in 50-rancher.yaml file to load my file:

"kubelet-arg": [
    "config=/tmp/kube-config.yaml"
  ],

the service restarts the logs do not show problems loading file

time="2023-05-10T16:16:00Z" level=info msg="Running kubelet --address=0.0.0.0 --alsologtostderr=false --anonymous-auth=false --authentication-token-webhook=true --authorization-mode=Webhook --cgroup-driver=systemd --client-ca-file=/var/lib/rancher/rke2/agent/client-ca.crt --cloud-provider=external --cluster-dns=10.43.0.10 --cluster-domain=cluster.local --config=/tmp/kube-config.yaml --container-runtime-endpoint=unix:///run/k3s/containerd/containerd.sock --containerd=/run/k3s/containerd/containerd.sock --eviction-hard=imagefs.available<5%,nodefs.available<5% --eviction-minimum-reclaim=imagefs.available=10%,nodefs.available=10% --fail-swap-on=false --healthz-bind-address=127.0.0.1 --hostname-override=dev-master-3 --kubeconfig=/var/lib/rancher/rke2/agent/kubelet.kubeconfig --log-file=/var/lib/rancher/rke2/agent/logs/kubelet.log --log-file-max-size=50 --logtostderr=false --node-ip=10.0.0.13 --node-labels=<http://cattle.io/os=linux,rke.cattle.io/machine=af095b89-819a-4320-9e60-9ac1d3d2f824,topology.kubernetes.io/region=hel1,topology.kubernetes.io/zone=hel1-dc2|cattle.io/os=linux,rke.cattle.io/machine=af095b89-819a-4320-9e60-9ac1d3d2f824,topology.kubernetes.io/region=hel1,topology.kubernetes.io/zone=hel1-dc2> --pod-infra-container-image=<http://index.docker.io/rancher/pause:3.6|index.docker.io/rancher/pause:3.6> --pod-manifest-path=/var/lib/rancher/rke2/agent/pod-manifests --read-only-port=0 --register-with-taints=<http://node-role.kubernetes.io/control-plane:NoSchedule,node-role.kubernetes.io/etcd:NoExecute|node-role.kubernetes.io/control-plane:NoSchedule,node-role.kubernetes.io/etcd:NoExecute> --resolv-conf=/run/systemd/resolve/resolv.conf --serialize-image-pulls=false --stderrthreshold=FATAL --tls-cert-file=/var/lib/rancher/rke2/agent/serving-kubelet.crt --tls-private-key-file=/var/lib/rancher/rke2/agent/serving-kubelet.key"

but from above I still see that it doe not take my config as

--eviction-hard=imagefs.available<5%,nodefs.available<5%

and I still seeing

Flag --eviction-hard has been deprecated, This parameter should be set via the config file specified by the Kubelet's --config flag. See <https://kubernetes.io/docs/tasks/administer-cluster/kubelet-config-file/> for more information.

Any ideas what is missing?

Hi Andrius, did you find the solution? Struggling with the same

I got it running. See this https://github.com/rancher/rke2/issues/4200