# k3s


10/31/2022, 10:10 AM
just upgraded from 1.24 to 1.25 and now it seems that (depends on what depends on what, but i assume tigera would need to come up first) tiger/calico does not come up and the ingress. I use the vanilla/public helm charts for both, so i boostrap k3s with:
Copy code
curl -fL <> | K3S_KUBECONFIG_MODE="640" INSTALL_K3S_EXEC="server --disable-kube-proxy --disable=servicelb --disable=servicelb --disable-network-policy --flannel-backend=none --disable traefik" sh
And then install the tigera helm chart (using v3.24.3 right now) and the nginx ingress helm chart (using 4.2.5).
I have looked at but was not able to find anything specific here. Is tigere 1.24 suppose to work with k3s 1.25 / k8s 1.25?
assuming what i see on it should actually work
Looking at the calico logs
Copy code
2022-10-31 10:16:15.923 [ERROR][7053] felix/discovery.go 153: Didn't find any ready Typha instances.
2022-10-31 10:16:15.923 [ERROR][7053] felix/daemon.go 323: Typha discovery enabled but discovery failed. error=Kubernetes service missing IP or port
and on the operator
Copy code
2022/10/31 10:16:09 [INFO] Version: v1.28.3
2022/10/31 10:16:09 [INFO] Go Version: go1.18.7b7
2022/10/31 10:16:09 [INFO] Go OS/Arch: linux/amd64
2022/10/31 10:16:39 [ERROR] Get "<>": dial tcp i/o timeout
i guess this rather means that the nginx ingress is the cause
All i see there is
Copy code
NGINX Ingress controller
  Release:       v1.3.1
  Build:         92534fa2ae799b502882c8684db13a25cde68155
  Repository:    <>
  nginx version: nginx/1.19.10


W1031 10:16:54.938970       7 client_config.go:617] Neither --kubeconfig nor --master was specified.  Using the inClusterConfig.  This might not work.
I1031 10:16:54.939191       7 main.go:209] "Creating API client" host="<>"
donwgrading to 1.24.7 fixes the issue, created a bug report


10/31/2022, 4:11 PM
I don't believe this is a k3s issue. Please check that you're configuring calico properly.
Other than the Kubernetes versions there are not any major differences between the two releases.


10/31/2022, 5:51 PM
i would assume it is eBPF (again and again)
At some point i ask myself is BPF (on calico) is actualy run by anybody in prod ;/