WireGuard 系列文章（九）：基于 K3S+WireGuard+Kilo 搭建跨多云的统一 K8S 集群

东风微鸣

发布于 2022-04-22 14:12:17

3.3K0

发布于 2022-04-22 14:12:17

文章被收录于专栏：东风微鸣技术博客东风微鸣技术博客

系列文章前情提要：

1.WireGuard 系列文章（一）：什么是 V**[1]

2.WireGuard 系列文章（二）：WireGuard 简介 - 快速、现代、安全的 V** 隧道[2]

3.WireGuard 系列文章（三）：WireGuard 安装[3]

4.WireGuard 系列文章（四）：WireGuard 快速上手[4]

5.WireGuard 系列文章（五）：Netmaker 简介-创建和管理 WireGuard 网络的平台[5]

6.WireGuard 系列文章（六）：Netmaker 安装[6]

7.WireGuard 系列文章（七）：使用 WireGuard 和 Netmaker 创建 Full Mesh 网络[7]

8.WireGuard 系列文章（八）：基于 WireGuard 的 K8S CNI Kilo 简介[8]

🛠️ 实战环节！基于 K3S + WireGuard + Kilo 搭建跨多云的统一 K8S 集群。💪💪💪

步骤

1. 前提

1.1 跨云的多台云主机

准备至少 2 台不同公有云的云主机（配置最低 1C1G 就能跑得动），这里准备了 6 台，主机名要求各不相同，分别是：

1.天翼云：ty1（K3S Server）

2.阿里云：ali（K3S Agent）

3.华为云：hw1（K3S Agent）

4.百度云：bd1 和 bd2（K3S Agent）

5.腾讯云：tx1（K3S Agent）

1.2 操作系统

操作系统：推荐 ubuntu 20.04（截止 2022/1/22），因为可以安装 WireGuard 方便。

1.3 已安装 WireGuard

且已安装 WireGuard，安装过程见👉️这里[9]

1.4 网络

协议	端口	源	描述
TCP	6443	K3s agent 节点	Kubernetes API Server
UDP	51820	K3s server 和 agent 节点	Kilo 网络间通信
TCP	10250	K3s server 和 agent 节点	Kubelet metrics
TCP	2379-2380	K3s server 节点	只有嵌入式 etcd 高可用才需要

通常情况下，所有出站流量都是允许的。

1.5 确保云主机有公网 IP

每个 location 至少有一个节点的 IP 地址与其他位置路由可达。

如果 location 在不同的云或私有网络中，那么这必须是一个公共 IP 地址。如果该 IP 地址没有在节点的以太网设备上自动配置，可以使用 kilo.squat.ai/force-endpoint[10] 注释方式手动指定。

1.6 云主机启用 IPv4 Forwarding

在 /etc/sysctl.conf 有这么一行：

net.ipv4.ip_forward=1

并生效：systel -p

1.7 （可选）配置镜像仓库配置

参考这里👉️：私有镜像仓库配置参考[11] 主要是为了加速 pull 镜像。

2. 安装 K3S Server

备注： K3S Server 可以简单的类比为 K8S 的 master。

2.1 一键安装 K3S Server

细节省略，直接通过 cnrancher 的安装脚本来一键安装：

curl -sfL http://rancher-mirror.cnrancher.com/k3s/k3s-install.sh | INSTALL_K3S_MIRROR=cn K3S_CLUSTER_INIT=true INSTALL_K3S_EXEC="--tls-san {{ server_public_ip }} --node-external-ip {{ master_ip }}  --flannel-backend none --kube-proxy-arg metrics-bind-address=0.0.0.0 --kube-apiserver-arg feature-gates=EphemeralContainers=true" sh -s -

简要说明：

1.K3S_CLUSTER_INIT=true: 集群模式，会安装内置的 etcd，而不是 sqlite3；

2.--tls-san {{ server_public_ip }}: {{ server_public_ip }} 改为你自己指定的 k3s server 公网 IP，--tls-san 这个选项在 TLS 证书中增加了一个额外的主机名或 IP 作为备用名称，如果你想通过 IP 和主机名访问，可以多次指定。

3.--node-external-ip {{ server_public_ip }}：指定 node 的公网 IP

4.--flannel-backend none：K3S 默认网络插件为 flannel，这里意思是不使用 flannel，后面单独安装 Kilo。

5.--kube-proxy-arg metrics-bind-address=0.0.0.0: kube-proxy 参数。

6.（可选）：--kube-apiserver-arg feature-gates=EphemeralContainers=true 启用 feature-gates：EphemeralContainers，方便在不重启 pod 的情况下附加 sidecar 进行调试。

❌ 错误： 1.如果不指定 --tls-san {{ server_public_ip }}，可能会导致 kubectl 无法通过 server_public_ip 来访问集群。
2.如果不指定 --node-external-ip {{ server_public_ip }}，可能会导致位于其他云的 k3s agent 无法连接到 K3S Server 的 API。3.如果不指定 --kube-proxy-arg metrics-bind-address=0.0.0.0，可能导致无法获取到 Metrics。

2.2 查看安装结果

具体如下：

❯ systemctl status k3s.service
● k3s.service - Lightweight Kubernetes
     Loaded: loaded (/etc/systemd/system/k3s.service; enabled; vendor preset: enabled)
     Active: active (running) since Sat 2022-01-22 16:27:14 CST; 4h 5min ago
       Docs: https://k3s.io
    Process: 5757 ExecStartPre=/sbin/modprobe br_netfilter (code=exited, status=0/SUCCESS)
    Process: 5758 ExecStartPre=/sbin/modprobe overlay (code=exited, status=0/SUCCESS)
   Main PID: 5759 (k3s-server)
      Tasks: 49
     Memory: 926.0M
     CGroup: /system.slice/k3s.service
             ├─ 5759 /usr/local/bin/k3s server
             ├─ 5774 containerd
             ├─18561 /var/lib/rancher/k3s/data/2e877cf4762c3c7df37cc556de3e08890fbf450914bb3ec042ad4f36b5a2413a/bin/containerd-shim-runc-v2 -namespace k>
             ├─18579 /pause
             └─18745 /opt/bin/kg --kubeconfig=/etc/kubernetes/kubeconfig --hostname=ty1.k3s

Jan 22 17:28:41 ty1.k3s k3s[5759]: I0122 17:28:41.435542    5759 reconciler.go:319] "Volume detached for volume...
...

❯ journalctl -f -b -u k3s.service
-- Logs begin at Fri 2021-03-26 09:47:06 CST, end at Sat 2022-01-22 20:35:12 CST. --
Jan 22 16:20:21 ty1.k3s systemd[1]: Starting Lightweight Kubernetes...
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Starting k3s v1.22.5+k3s1 (405bf79d)"
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Configuring sqlite3 database connection pooling: maxIdleConns=2, max>
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Configuring database table schema and indexes, this may take a momen>
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Database tables and indexes are up to date"
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Kine available at unix://kine.sock"
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Reconciling bootstrap data between datastore and disk"
Jan 22 16:20:22 ty1.k3s k3s[1660]: time="2022-01-22T16:20:22+08:00" level=info msg="Running kube-apiserver --advertise-address=x.x.x.x --adverti>
Jan 22 16:20:22 ty1.k3s k3s[1660]: Flag --insecure-port has been deprecated, This flag has no effect now and will be removed in v1.24.
Jan 22 16:20:22 ty1.k3s k3s[1660]: I0122 16:20:22.307328    1660 server.go:581] external host was not specified, using x.x.x.x
Jan 22 16:20:22 ty1.k3s k3s[1660]: I0122 16:20:22.308688    1660 server.go:175] Version: v1.22.5+k3s1

❯ k3s kubectl cluster-info
Kubernetes control plane is running at https://127.0.0.1:6443
CoreDNS is running at https://127.0.0.1:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
Metrics-server is running at https://127.0.0.1:6443/api/v1/namespaces/kube-system/services/https:metrics-server:/proxy

❯ k3s kubectl get node
NAME       STATUS   ROLES                  AGE     VERSION
ty1.k3s    NotReady    control-plane,master   4h10m   v1.22.5+k3s1

⚠️ 注意：此时 K3S Server 的状态是 NotReady，这是正常状况，因为还没有安装 CNI 网络插件。

3. 安装 Kilo

3.1 指定 K3S Server 拓扑

我的 K3S Server 在天翼云上，指定拓扑：

k3s kubectl annotate node ty1.k3s kilo.squat.ai/location="ctyun"
k3s kubectl annotate node ty1.k3s kilo.squat.ai/force-endpoint=x.x.x.x:51820
k3s kubectl annotate node ty1.k3s kilo.squat.ai/persistent-keepalive=20

解释在这里👉️：Kilo Annotions - location[12]

3.2 安装 Kilo❗️

通过在集群中部署一个 DaemonSet 来安装 Kilo。

kubectl apply -f https://gitee.com/mirrors/squat/raw/main/manifests/crds.yaml
kubectl apply -f https://gitee.com/mirrors/squat/raw/main/manifests/kilo-k3s.yaml

ℹ️ 备注：上面的地址是 Kilo 的 Gitee mirror 仓库。

详细说明：

1.crds.yaml 安装了 peer.kilo.squat.ai，用于通过该 CRD 配置 WireGuard Peer。

2.kilo-k3s.yaml 安装了 Kilo CNI，包括：

1.ConfigMap1.cni-conf.json

2.kilo-scripts

2.ServiceAccount3.ClusterRole4.ClusterRoleBinding5.DaemonSet: 运行在所有 Node 上

3.其中，WireGuard 的配置：

1.ini 格式的 WireGuard 配置位于：/var/lib/kilo/conf

2.WireGuard 的 key 位于：/var/lib/kilo/key

3.3 验证

1.kube-system ns 中，已创建 kilo DaemonSet，且所有 Pod 已出于 Running 状态2.K3S Server Node 状态从 NotReady 变为 Ready3.K3S Server Node 上，已经有 kilo 的相关 annotations, 示例如下：

apiVersion: v1
kind: Node
metadata:
  name: ty1.k3s
  labels:
    beta.kubernetes.io/arch: amd64
    ...
  annotations:
    k3s.io/external-ip: x.x.x.x
    k3s.io/hostname: ty1.k3s
    k3s.io/internal-ip: 192.168.1.226
    k3s.io/node-args: >-
      ["server","--tls-san","x.x.x.x","--node-external-ip","x.x.x.x","--flannel-backend","none","--kube-proxy-arg","metrics-bind-address=0.0.0.0","--kube-apiserver-arg","feature-gates=EphemeralContainers=true"]
    kilo.squat.ai/endpoint: x.x.x.x:51820
    kilo.squat.ai/force-endpoint: x.x.x.x:51820
    kilo.squat.ai/granularity: location
    kilo.squat.ai/internal-ip: 192.168.1.226/24
    kilo.squat.ai/key: zCiXXXXXXXXXXXXXXXXXXXXXXXXXXXXQTL9CEc=
    kilo.squat.ai/last-seen: '1642856638'
    kilo.squat.ai/location: ctyun
    kilo.squat.ai/persistent-keepalive: '20'
    kilo.squat.ai/wireguard-ip: 10.4.0.3/16
...

4. 安装 K3S Agent

4.1 一键安装 K3S Agent

curl -sfL http://rancher-mirror.cnrancher.com/k3s/k3s-install.sh | INSTALL_K3S_MIRROR=cn K3S_TOKEN={{ token }} K3S_URL=https://{{ server_public_ip }}:6443 sh -s - --node-external-ip {{ node_public_ip }} --kube-proxy-arg "metrics-bind-address=0.0.0.0"

4.2 等待 K3S Agent 加入集群中

等待 K3S Agent 加入集群中，状态 NotReady 没关系：

❯ systemctl status k3s-agent.service
● k3s-agent.service - Lightweight Kubernetes
     Loaded: loaded (/etc/systemd/system/k3s-agent.service; enabled; vendor preset: enabled)
     Active: active (running) since Sat 2022-01-22 16:27:35 CST; 4h 44min ago
       Docs: https://k3s.io
    Process: 4079 ExecStartPre=/sbin/modprobe br_netfilter (code=exited, status=0/SUCCESS)
    Process: 4080 ExecStartPre=/sbin/modprobe overlay (code=exited, status=0/SUCCESS)
   Main PID: 4081 (k3s-agent)
      Tasks: 63
     Memory: 126.9M
     CGroup: /system.slice/k3s-agent.service
             ├─4081 /usr/local/bin/k3s agent
             ├─4106 containerd
             ├─5285 /var/lib/rancher/k3s/data/
             ...

-- Logs begin at Sat 2021-11-06 14:00:29 CST, end at Sat 2022-01-22 21:10:33 CST. --
Jan 22 16:27:35 ali1.k3s systemd[1]: Starting Lightweight Kubernetes...
Jan 22 16:27:35 ali1.k3s systemd[1]: Started Lightweight Kubernetes.
Jan 22 16:27:35 ali1.k3s k3s[4081]: time="2022-01-22T16:27:35+08:00" level=info msg="Starting k3s agent v1.22.5+k3s1 (405bf79d)"
Jan 22 16:27:35 ali1.k3s k3s[4081]: time="2022-01-22T16:27:35+08:00" level=info msg="Running load balancer 127.0.0.1:6444 -> [192.168.1.226:6443 140.246.255.203:6443]"
Jan 22 16:27:55 ali1.k3s k3s[4081]: time="2022-01-22T16:27:55+08:00" level=error msg="failed to get CA certs: Get \"https://127.0.0.1:6444/cacerts\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Module overlay was already loaded"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Module nf_conntrack was already loaded"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Module br_netfilter was already loaded"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Module iptable_nat was already loaded"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Using private registry config file at /etc/rancher/k3s/registries.yaml"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Logging containerd to /var/lib/rancher/k3s/agent/containerd/containerd.log"
Jan 22 16:28:01 ali1.k3s k3s[4081]: time="2022-01-22T16:28:01+08:00" level=info msg="Running containerd -c /var/lib/rancher/k3s/agent/etc/containerd/config.toml -a /run/k3s/containerd/containerd.sock --state /run/k3s/containerd --root /var/lib/rancher/k3s/agent/containerd"
Jan 22 16:28:02 ali1.k3s k3s[4081]: time="2022-01-22T16:28:02+08:00" level=info msg="Containerd is now running"
Jan 22 16:28:02 ali1.k3s k3s[4081]: time="2022-01-22T16:28:02+08:00" level=info msg="Updating load balancer server addresses -> [140.246.255.203:6443]"
Jan 22 16:28:02 ali1.k3s k3s[4081]: time="2022-01-22T16:28:02+08:00" level=info msg="Connecting to proxy" url="wss://140.246.255.203:6443/v1-k3s/connect"
Jan 22 16:28:02 ali1.k3s k3s[4081]: time="2022-01-22T16:28:02+08:00" level=info msg="Running kubelet --address=0.0.0.0 --anonymous-auth=false --authentication-token-webhook=true --authorization-mode=Webhook --cgroup-driver=cgroupfs --client-ca-file=/var/lib/rancher/k3s/agent/cli>
Jan 22 16:28:02 ali1.k3s k3s[4081]: Flag --cloud-provider has been deprecated, will be removed in 1.23, in favor of removing cloud provider code from Kubelet.
Jan 22 16:28:02 ali1.k3s k3s[4081]: Flag --containerd has been deprecated, This is a cadvisor flag that was mistakenly registered with the Kubelet. Due to legacy concerns, it will follow the standard CLI deprecation timeline before being removed.
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.742554    4081 server.go:436] "Kubelet version" kubeletVersion="v1.22.5+k3s1"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.798379    4081 dynamic_cafile_content.go:155] "Starting controller" name="client-ca-bundle::/var/lib/rancher/k3s/agent/client-ca.crt"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.891854    4081 server.go:687] "--cgroups-per-qos enabled, but --cgroup-root was not specified.  defaulting to /"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.892197    4081 container_manager_linux.go:280] "Container manager verified user specified cgroup-root exists" cgroupRoot=[]
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.892310    4081 container_manager_linux.go:285] "Creating Container Manager object based on Node Config" nodeConfig={RuntimeCgroupsName: SystemCgroupsName: KubeletCgroupsName: ContainerRuntime:remote CgroupsPerQOS:true CgroupRoot>
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.892345    4081 topology_manager.go:133] "Creating topology manager with policy per scope" topologyPolicyName="none" topologyScopeName="container"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.892357    4081 container_manager_linux.go:320] "Creating device plugin manager" devicePluginEnabled=true
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.892403    4081 state_mem.go:36] "Initialized new in-memory state store"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.893507    4081 kubelet.go:418] "Attempting to sync node with API server"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.893533    4081 kubelet.go:279] "Adding static pod path" path="/var/lib/rancher/k3s/agent/pod-manifests"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.893577    4081 kubelet.go:290] "Adding apiserver pod source"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.893601    4081 apiserver.go:42] "Waiting for node sync before watching apiserver pods"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.897574    4081 kuberuntime_manager.go:245] "Container runtime initialized" containerRuntime="containerd" version="v1.5.8-k3s1" apiVersion="v1alpha2"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.898404    4081 server.go:1213] "Started kubelet"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.901202    4081 fs_resource_analyzer.go:67] "Starting FS ResourceAnalyzer"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.903670    4081 server.go:149] "Starting to listen" address="0.0.0.0" port=10250
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.904727    4081 server.go:409] "Adding debug handlers to kubelet server"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.909003    4081 volume_manager.go:291] "Starting Kubelet Volume Manager"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.909628    4081 desired_state_of_world_populator.go:146] "Desired state populator starts to run"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.949667    4081 kubelet_network_linux.go:56] "Initialized protocol iptables rules." protocol=IPv4
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.971847    4081 kubelet_network_linux.go:56] "Initialized protocol iptables rules." protocol=IPv6
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.972256    4081 status_manager.go:158] "Starting to sync pod status with apiserver"
Jan 22 16:28:02 ali1.k3s k3s[4081]: I0122 16:28:02.972423    4081 kubelet.go:1967] "Starting kubelet main sync loop"
...
Jan 22 16:28:03 ali1.k3s k3s[4081]: I0122 16:28:03.331465    4081 kubelet_node_status.go:74] "Successfully registered node" node="ali1.k3s"

❯ k3s kubectl get node
NAME       STATUS   ROLES                  AGE     VERSION
ali1.k3s   NotReady    worker                 4h41m   v1.22.5+k3s1
bd2.k3s    NotReady    worker                 4h40m   v1.22.5+k3s1
ty1.k3s    Ready    control-plane,master   4h42m   v1.22.5+k3s1
tx1.k3s    NotReady    worker                 4h41m   v1.22.5+k3s1
hw1.k3s    NotReady    worker                 4h41m   v1.22.5+k3s1
bd1.k3s    NotReady    worker                 4h40m   v1.22.5+k3s1

4.3 指定 K3S Agent 拓扑

# ali1
k3s kubectl annotate node ali1.k3s kilo.squat.ai/location="aliyun"
k3s kubectl annotate node ali1.k3s kilo.squat.ai/force-endpoint={{ ali1_public_ip }}:51820
k3s kubectl annotate node ali1.k3s kilo.squat.ai/persistent-keepalive=20

# hw1
k3s kubectl annotate node hw1.k3s kilo.squat.ai/location="huaweicloud"
k3s kubectl annotate node hw1.k3s kilo.squat.ai/force-endpoint={{ hw1_public_ip }}:51820
k3s kubectl annotate node hw1.k3s kilo.squat.ai/persistent-keepalive=20

# bd1
k3s kubectl annotate node bd1.k3s kilo.squat.ai/location="baidu"
k3s kubectl annotate node bd1.k3s kilo.squat.ai/force-endpoint={{ bd1_public_ip }}:51820
k3s kubectl annotate node bd1.k3s kilo.squat.ai/persistent-keepalive=20

# bd2
k3s kubectl annotate node bd2.k3s kilo.squat.ai/location="baidu"
k3s kubectl annotate node bd2.k3s kilo.squat.ai/force-endpoint={{ bd2_public_ip }}:51820
k3s kubectl annotate node bd2.k3s kilo.squat.ai/persistent-keepalive=20

# tx1
k3s kubectl annotate node tx1.k3s kilo.squat.ai/location="tencentcloud"
k3s kubectl annotate node tx1.k3s kilo.squat.ai/force-endpoint={{ tx2_public_ip }}:51820
k3s kubectl annotate node tx1.k3s kilo.squat.ai/persistent-keepalive=20

等待所有的节点状态 Ready.

5. 验证

5.1 验证网络通断

部署 1 个 busybox 的 DaemonSet，使其落在每一个 Node 上。验证网络流向：

进入其中一个 pod，ping 其他 pod，可以互相 ping 通，示例如下：

/ # ping 10.42.2.7
PING 10.42.2.7 (10.42.2.7): 56 data bytes
64 bytes from 10.42.2.7: seq=0 ttl=62 time=6.604 ms
64 bytes from 10.42.2.7: seq=1 ttl=62 time=6.520 ms
64 bytes from 10.42.2.7: seq=2 ttl=62 time=6.412 ms
64 bytes from 10.42.2.7: seq=3 ttl=62 time=6.430 ms
64 bytes from 10.42.2.7: seq=4 ttl=62 time=6.487 ms
^C
--- 10.42.2.7 ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max = 6.412/6.490/6.604 ms
/ # ping 10.42.1.3
PING 10.42.1.3 (10.42.1.3): 56 data bytes
64 bytes from 10.42.1.3: seq=0 ttl=62 time=7.426 ms
64 bytes from 10.42.1.3: seq=1 ttl=62 time=7.123 ms
64 bytes from 10.42.1.3: seq=2 ttl=62 time=7.109 ms
64 bytes from 10.42.1.3: seq=3 ttl=62 time=7.129 ms
^C
--- 10.42.1.3 ping statistics ---
4 packets transmitted, 4 packets received, 0% packet loss
round-trip min/avg/max = 7.109/7.196/7.426 m

ping 其他 node 的 WireGuard 地址，也可以 ping 通，如下：

/ # ping 10.4.0.1
PING 10.4.0.1 (10.4.0.1): 56 data bytes
64 bytes from 10.4.0.1: seq=0 ttl=64 time=0.077 ms
64 bytes from 10.4.0.1: seq=1 ttl=64 time=0.099 ms
^C
--- 10.4.0.1 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 0.077/0.088/0.099 ms
/ # ping 10.4.0.2
PING 10.4.0.2 (10.4.0.2): 56 data bytes
64 bytes from 10.4.0.2: seq=0 ttl=63 time=29.000 ms
64 bytes from 10.4.0.2: seq=1 ttl=63 time=28.939 ms
^C
--- 10.4.0.2 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 28.939/28.969/29.000 ms

ping 其他 node 的默认内网地址，也可以 ping 通，如下：

/ # ping 172.17.0.3
PING 172.17.0.3 (172.17.0.3): 56 data bytes
64 bytes from 172.17.0.3: seq=0 ttl=63 time=6.327 ms
64 bytes from 172.17.0.3: seq=1 ttl=63 time=6.350 ms
^C
--- 172.17.0.3 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 6.327/6.338/6.350 ms
/ # ping 192.168.64.4
PING 192.168.64.4 (192.168.64.4): 56 data bytes
64 bytes from 192.168.64.4: seq=0 ttl=63 time=29.261 ms
64 bytes from 192.168.64.4: seq=1 ttl=63 time=29.015 ms
^C
--- 192.168.64.4 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 29.015/29.138/29.261 ms

证明网络（pod-pod，pod-node）全通。

5.2 ℹ️ 详细说明

5.2.1 网卡

Kilo 建立了 3 个网卡，分别为：

1.kilo0：WireGuard V** 网络，用以 Node 间组建 V** 内网；（bd2 上没有该网卡，因为和 bd1 同属一个 VPC 内网，内网内连接默认不通过 WireGuard 加密，只有出 VPC 内网才会走 WireGuard 加密）

2.kube-bridge: 桥接网络，使得 Pod 的网卡与云主机的网卡进行连接，以此实现 Pod 通过 WireGuard V** 网络共享通信；

1.tunl0：Bridge 模式下，多主机网络通信需要额外配置主机路由，或使用 overlay 网络。通过 Kilo 来自动配置。比如 overlay 情况下的网络结构为：

典型 Bridge Overlay 网络结构

其中，kube-bridge 配置如下：

{
    "cniVersion": "0.3.1",
    "name": "kilo",
    "plugins": [
        {
            "name": "kubernetes",
            "type": "bridge",
            "bridge": "kube-bridge",
            "isDefaultGateway": true,
            "forceAddress": true,
            "mtu": 1420,
            "ipam": {
                "type": "host-local"
            }
        },
        {
            "type": "portmap",
            "snat": true,
            "capabilities": {
                "portMappings": true
            }
        }
    ]
}

5.2.2 CIDR

如下：

云主机	自带内网 IP	WireGuard V** IP	Pod CIDR
ty1	192.168.1.226	10.4.0.3/16	10.42.0.0/24
ali1	172.21.143.136	10.4.0.1/16	10.42.3.0/24
hw1	192.168.7.226	10.4.0.4/16	10.42.1.0/24
bd1	192.168.64.4	10.4.0.2/16	10.42.4.0/24
bd2	192.168.64.5	无	10.42.5.0/24
tx1	172.17.0.3	10.4.0.5/16	10.42.2.0/24

ℹ️ 备注：云主机的自带内网 IP 没有经过任何特殊设置，都是公有云自动生成的。

5.2.3 路由表

再看路由表，以 ty1 为例，如下：

❯ route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         192.168.1.1     0.0.0.0         UG    100    0        0 eth0
10.4.0.0        0.0.0.0         255.255.0.0     U     0      0        0 kilo0
10.42.0.0       0.0.0.0         255.255.255.0   U     0      0        0 kube-bridge
10.42.1.0       10.4.0.4        255.255.255.0   UG    0      0        0 kilo0
10.42.2.0       10.4.0.5        255.255.255.0   UG    0      0        0 kilo0
10.42.3.0       10.4.0.1        255.255.255.0   UG    0      0        0 kilo0
10.42.4.0       10.4.0.2        255.255.255.0   UG    0      0        0 kilo0
10.42.5.0       10.4.0.2        255.255.255.0   UG    0      0        0 kilo0
172.17.0.3      10.4.0.5        255.255.255.255 UGH   0      0        0 kilo0
172.21.143.136  10.4.0.1        255.255.255.255 UGH   0      0        0 kilo0
192.168.7.226   10.4.0.4        255.255.255.255 UGH   0      0        0 kilo0
192.168.64.4    10.4.0.2        255.255.255.255 UGH   0      0        0 kilo0
192.168.64.5    10.4.0.2        255.255.255.255 UGH   0      0        0 kilo0

bd1 的如下：

❯ route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
...
10.42.5.0       192.168.64.5    255.255.255.0   UG    0      0        0 tunl0
...

bd2 的如下：

❯ route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         192.168.64.1    0.0.0.0         UG    100    0        0 eth0
10.4.0.1        192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
10.4.0.2        192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
10.4.0.3        192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
10.4.0.4        192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
10.4.0.5        192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
10.42.0.0       192.168.64.4    255.255.255.0   UG    0      0        0 tunl0
10.42.1.0       192.168.64.4    255.255.255.0   UG    0      0        0 tunl0
10.42.2.0       192.168.64.4    255.255.255.0   UG    0      0        0 tunl0
10.42.3.0       192.168.64.4    255.255.255.0   UG    0      0        0 tunl0
10.42.4.0       192.168.64.4    255.255.255.0   UG    0      0        0 tunl0
10.42.5.0       0.0.0.0         255.255.255.0   U     0      0        0 kube-bridge
169.254.169.254 192.168.64.2    255.255.255.255 UGH   100    0        0 eth0
172.17.0.3      192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
172.21.143.136  192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
192.168.1.226   192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
192.168.7.226   192.168.64.4    255.255.255.255 UGH   0      0        0 tunl0
192.168.64.0    0.0.0.0         255.255.240.0   U     0      0        0 eth0

5.2.4 总结

1.这台访问 10.4.0.0/16（也就是 WireGuard 网段），都走 kilo0 网卡（就是 WireGuard 网卡）；

1.云主机间通过 WireGuard V** IP 互相访问；

2.ty1 上 pod 互访（就是访问 10.42.0.0/24 网段），走 kube-bridge 网桥，

3.ty1 访问其他 Node 的 Pod, 通过 kilo0:

1.访问 10.42.1.0/24(hw1 上的 pod)，SNAT 走 hw1 的 WireGuard V** IP：10.4.0.4

2.访问 10.42.2.0/24(tx1 上的 pod)，SNAT 走 tx1 的 WireGuard V** IP：10.4.0.5

3.访问 10.42.3.0/24(ali1 上的 pod)，SNAT 走 ali1 的 WireGuard V** IP：10.4.0.1

4.访问 10.42.4.0/24(bd1 上的 pod) 和 10.42.5.0/24（bd2 上的 pod），SNAT 都走 bd1 的 WireGuard V** IP：10.4.0.2（因为 bd1 和 bd2 位于同一 VPC 内网，且只有 bd1 上有 WireGuard V** IP，bd2 上没有）

4.ty1 通过云主机自带的内网 IP 访问其他 Node，也是通过 kilo0 SNAT 后走对应的 WireGuard V** IP5.bd1 访问 10.42.5.0/24（bd2 上的 pod）, 通过 tunl0.6.bd2 上，就只有 kube-bridge 和 tunl0

1.同一台机器 pod 互访，kube-bridge

2.访问其他机器 pod，走 tunl0 通过 VPC 内网 IP（192.168.64.4）到 bd，再由 bd1 转发出去。

🗒️ 再凝练下：

1.同一 Node 里，走网桥；2.同一 location，走自带的内网 IP（VPC 网络）；3.不同 location，走 WireGuard VP**；

`kgctl`

Kilo 提供了一个用于检查集群和与集群交互的命令行工具：kgctl。

这个工具可以用来理解一个网格的拓扑结构，获取对等点的 WireGuard 配置，或者绘制一个集群图。

kgctl需要通过设置KUBECONFIG环境变量或提供--kubeconfig标志来提供 Kubernetes 配置文件。

安装

kgctl二进制文件是为 Linux、macOS 和 Windows 自动编译的，每个版本的 Kilo 都可以从 GitHub 发布页面[13] 下载。

命令

命令	语法	描述
graph[14]	kgctl graph [flags]	以 GraphViz 格式生成表示集群拓扑的图形。
showconf[15]	kgctl showconf ( node \| peer ) NAME [flags]	显示网格中节点或对等体的 WireGuard 配置。

graph

⚠️ 注意：需要安装 circo，否则无法生成拓扑图。安装命令：sudo apt install circos graphviz -y

graph命令以 GraphViz 格式生成一个表示 Kilo 网格的图形。该图有助于理解或调试网络拓扑。

例子：

kgctl graph

这将产生一些 DOT 图描述语言的输出，例如：

digraph kilo {
    label="10.2.4.0/24";
    labelloc=t;
    outputorder=nodesfirst;
    overlap=false;
    "ip-10-0-6-7"->"ip-10-0-6-146"[ dir=both ];
    "ip-10-1-13-74"->"ip-10-1-20-76"[ dir=both ];
    "ip-10-0-6-7"->"ip-10-1-13-74"[ dir=both ];
    "ip-10-0-6-7"->"squat"[ dir=both, style=dashed ];
    "ip-10-1-13-74"->"squat"[ dir=both, style=dashed ];

# ...

}
;

要渲染图形，使用 GraphViz layout 工具之一，例如circo:

kgctl graph | circo -Tsvg > cluster.svg

showconf

showconf命令输出集群中某个节点或对等体的 WireGuard 配置，即该节点或对等体为了参与 mesh 需要在其本地的 WireGuard 接口上设置的配置。例子：

NODE=master # the name of a node
kgctl showconf node $NODE

这将产生一些 INI 格式的输出，例如。

[Interface]
ListenPort = 51820

[Peer]
AllowedIPs = 10.2.0.0/24, 10.1.13.74/32, 10.2.4.0/24, 10.1.20.76/32, 10.4.0.2/32
Endpoint = 3.120.246.76:51820
PersistentKeepalive = 0
PublicKey = IgDTEvasUvxisSAmfBKh8ngFmc2leZBvkRwYBhkybUg=

总结

WireGuard 配置

经过 Kilo 自动感知生成的 WireGuard 配置如下：

❯ kgctl showconf node ty1.k3s
[Interface]
ListenPort = 51820

[Peer]
AllowedIPs = 10.42.3.0/24, 172.21.143.136/32, 10.4.0.1/32
Endpoint = [{{ ali1_public_ip }}]:51820
PersistentKeepalive = 20
PublicKey = tscPxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx=

[Peer]
AllowedIPs = 10.42.4.0/24, 192.168.64.4/32, 10.42.5.0/24, 192.168.64.5/32, 10.4.0.2/32
Endpoint = [{{ bd1_public_ip }}]:51820
PersistentKeepalive = 20
PublicKey = 29khxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxrz8=

[Peer]
AllowedIPs = 10.42.1.0/24, 192.168.7.226/32, 10.4.0.4/32
Endpoint = [{{ hw1_public_ip }}]:51820
PersistentKeepalive = 20
PublicKey = B9JZe6X8+xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx=

[Peer]
AllowedIPs = 10.42.2.0/24, 172.17.0.3/32, 10.4.0.5/32
Endpoint = [{{ tx1_public_ip }}]:51820
PersistentKeepalive = 20
PublicKey = mn1rUiD+Zb3/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxU=

网络拓扑

如下图：

基于 K3S WireGuard Kilo 的多云统一 K8S 集群网络拓扑

网络流

1.同一 Node 里，走网桥；

2.同一 location，走自带的内网 IP（VPC 网络）；

3.不同 location，走 WireGuard V**；

🎉🎉🎉

References

[1] WireGuard 系列文章（一）：什么是 V**: http://ewhisper.cn/posts/32152/ [2] WireGuard 系列文章（二）：WireGuard 简介 - 快速、现代、安全的 V** 隧道: http://ewhisper.cn/posts/23685/ [3] WireGuard 系列文章（三）：WireGuard 安装: http://ewhisper.cn/posts/59274/ [4] WireGuard 系列文章（四）：WireGuard 快速上手: http://ewhisper.cn/posts/29790/ [5] WireGuard 系列文章（五）：Netmaker 简介-创建和管理 WireGuard 网络的平台: http://ewhisper.cn/posts/54509/ [6] WireGuard 系列文章（六）：Netmaker 安装: https://ewhisper.cn/posts/13793/ [7] WireGuard 系列文章（七）：使用 WireGuard 和 Netmaker 创建 Full Mesh 网络: https://ewhisper.cn/posts/24019/ [8] WireGuard 系列文章（八）：基于 WireGuard 的 K8S CNI Kilo 简介: https://ewhisper.cn/posts/3375/ [9] 这里: https://ewhisper.cn/posts/59274/ [10] kilo.squat.ai/force-endpoint: https://ewhisper.cn/posts/3375/#force-endpoint [11] 私有镜像仓库配置参考: https://docs.rancher.cn/docs/k3s/installation/private-registry/_index [12] Kilo Annotions - location: https://ewhisper.cn/posts/3375/#location [13] GitHub 发布页面: https://github.com/squat/kilo/releases/latest [14] graph: #graph [15] showconf: #showconf

本文参与腾讯云自媒体分享计划，分享自微信公众号。

原始发表：2022-01-23，如有侵权请联系 cloudcommunity@tencent.com 删除

tcp/ip

本文分享自东风微鸣技术博客微信公众号，前往查看

如有侵权，请联系 cloudcommunity@tencent.com 删除。

本文参与腾讯云自媒体分享计划，欢迎热爱写作的你一起参与！