Ubuntu22.04上的k3s网络问题:未能获得Kubernetes服务器版本:获取k3s i/o超时

Question

在从Ubuntu20.04LTS升级到Ubuntu22.04LTS之后，我目前正面临着k3s集群的问题。

例如，local-path-provisioner荚的日志：

$ kubectl logs -n kube-system local-path-provisioner-6c79684f77-l4cqp
time="2022-04-28T03:27:00Z" level=fatal msg="Error starting daemon: Cannot start Provisioner: failed to get Kubernetes server version: Get \"https://10.43.0.1:443/version?timeout=32s\": dial tcp 10.43.0.1:443: i/o timeout"

我尝试了以下操作：

• 禁用ipv6，如这里所述
• 禁用ufw防火墙
• 使用遗留iptables
• 向iptable添加内部通信规则，如下所示：

$ sudo iptables -A INPUT -s 10.42.0.0/16 -d <host_ip> -j ACCEPT

不过，coredns、local-path-provisioner和metrics-server的部署不会启动。当列出豆荚时，下面是输出：

$ kubectl get pods -n kube-system                                    
NAME                                      READY   STATUS             RESTARTS         AGE
cilium-64r4c                              1/1     Running            2 (18m ago)      174m
cilium-d8grw                              1/1     Running            2 (18m ago)      174m
cilium-g4gmf                              1/1     Running            2 (18m ago)      174m
cilium-h5j4h                              1/1     Running            2 (18m ago)      174m
cilium-n62nv                              1/1     Running            2 (18m ago)      174m
cilium-operator-76cff99967-6fgkv          1/1     Running            2 (18m ago)      174m
cilium-operator-76cff99967-pbr4l          1/1     Running            2 (18m ago)      174m
cilium-w4n6d                              1/1     Running            2 (18m ago)      174m
cilium-wgm7l                              1/1     Running            2 (18m ago)      174m
cilium-zqb6w                              1/1     Running            2 (18m ago)      174m
coredns-d76bd69b-bhgnl                    0/1     CrashLoopBackOff   44 (3m27s ago)   177m
hubble-relay-67f64789c7-cjzz9             0/1     CrashLoopBackOff   63 (4m15s ago)   174m
hubble-ui-794cd44b77-9vgbl                3/3     Running            6 (18m ago)      174m
local-path-provisioner-6c79684f77-l4cqp   0/1     CrashLoopBackOff   35 (3m53s ago)   177m
metrics-server-7cd5fcb6b7-v74rc           0/1     CrashLoopBackOff   42 (3m35s ago)   177m

任何帮助都是非常感谢的！谢谢

Answer 1

既然你用的是纤毛，我想你可能会遇到这样的问题：https://github.com/cilium/cilium/issues/10645

解决方法是确保将net.ipv4.conf.lxc*.rp_filter设置为0：

echo 'net.ipv4.conf.lxc*.rp_filter = 0' | sudo tee -a /etc/sysctl.d/90-override.conf
sudo systemctl start systemd-sysctl