Score:0

Custer does not start. etcd and kube-apiserver do not start. Kubelet starts

rs flag

After a blackout, the cluster does not start. 3 workers + 2 Control Planes. Installed with kubeadm. Version 1.27.1 containerd 1.6.22 Ubuntu 22.04.3 LTS It has been running fine for more than one year, and the last kube* upgrade was 2-3 months ago.

containerd starts. Kubelet starts, but none of the pods are launched, neither the etcd nor the kube-apiserver. So kubelet does not execute the /etc/kubernetes/manifests/.yaml I verified this point with crictl. The etcd and kube pods are dated 5 weeks ago, last time, I restarted the servers. And their logs refers to the last day they ran. Two control Planes have the same behavior, which is pretty suspicious. I activated the debug mode and verbosity in kubelet, containerd, pods, ... Nothing clear.

I thought it was related with the network, and names resolution, because I read that kubernetes over Ubuntu can have problems with resolv.conf. But I verified that /var/lib/kubelet/config.yaml points to the correct resolv.conf. resolvConf: /run/systemd/resolve/resolv.conf

Certificates are fine.

In /run/containerd/ there is not io.containerd.grpc.v1.cri, runc nor s directories.

I checked all the logs but I do not know how to go on. Any idea?

I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.