使用client-go实现自定义控制器
介绍
我们已经知道,Service对集群之外暴露服务的主要方式有两种:NodePort和LoadBalancer,但是这两种方式,都有一定的缺点:
- NodePort方式的缺点是会占用很多集群机器的端口,那么当集群服务变多的时候,这个缺点就愈发明显。
- LoadBalancer的缺点是每个Service都需要一个LB,浪费,麻烦,并且需要Kubernetes之外的设备的支持。
基于这种现状,Kubernetes提供了Ingress资源对象,Ingress只需要一个NodePort或者一个LB就可以满足暴露多个Service的需求。
客户端首先对 域名 执行 DNS 解析,得到 Ingress Controller 所在节点的 IP,然后客户端向 Ingress Controller 发送 HTTP 请求,然后根据 Ingress 对象里面的描述匹配域名,找到对应的 Service 对象,并获取关联的 Endpoints 列表,将客户端的请求转发给其中一个 Pod。
本文我们来使用client-go实现一个自定义控制器,通过判断service
的Annotations
属性是否包含ingress/http
字段,如果包含则创建ingress,如果不包含则不创建。而且如果存在ingress
则进行删除。
具体实现
首先我们创建项目。
$ mkdir ingress-manager && cd ingress-manager
$ go mod init ingress-manager
# 由于控制器部分的内容比较多,将它们单独放到pkg目录下
$ mkdir pkg
# 最终项目目录结构如下
.
├── go.mod
├── go.sum
├── main.go
└── pkg
└── controller.go
接着我们来实现controller部分:
package pkg
import (
"context"
apiCoreV1 "k8s.io/api/core/v1"
netV1 "k8s.io/api/networking/v1"
"k8s.io/apimachinery/pkg/api/errors"
metaV1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/util/runtime"
"k8s.io/apimachinery/pkg/util/wait"
informersCoreV1 "k8s.io/client-go/informers/core/v1"
informersNetV1 "k8s.io/client-go/informers/networking/v1"
"k8s.io/client-go/kubernetes"
coreV1 "k8s.io/client-go/listers/core/v1"
v1 "k8s.io/client-go/listers/networking/v1"
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/util/workqueue"
"reflect"
"time"
)
const (
workNum = 5 // 工作的节点数
maxRetry = 10 // 最大重试次数
)
// 定义控制器
type Controller struct {
client kubernetes.Interface
ingressLister v1.IngressLister
serviceLister coreV1.ServiceLister
queue workqueue.RateLimitingInterface
}
// 初始化控制器
func NewController(client kubernetes.Interface, serviceInformer informersCoreV1.ServiceInformer, ingressInformer informersNetV1.IngressInformer) Controller {
c := Controller{
client: client,
ingressLister: ingressInformer.Lister(),
serviceLister: serviceInformer.Lister(),
queue: workqueue.NewNamedRateLimitingQueue(workqueue.DefaultControllerRateLimiter(), "ingressManager"),
}
// 添加事件处理函数
serviceInformer.Informer().AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: c.addService,
UpdateFunc: c.updateService,
})
ingressInformer.Informer().AddEventHandler(cache.ResourceEventHandlerFuncs{
DeleteFunc: c.deleteIngress,
})
return c
}
// 入队
func (c *Controller) enqueue(obj interface{}) {
key, err := cache.MetaNamespaceKeyFunc(obj)
if err != nil {
runtime.HandleError(err)
}
c.queue.Add(key)
}
func (c *Controller) addService(obj interface{}) {
c.enqueue(obj)
}
func (c *Controller) updateService(oldObj, newObj interface{}) {
// todo 比较annotation
// 这里只是比较了对象是否相同,如果相同,直接返回
if reflect.DeepEqual(oldObj, newObj) {
return
}
c.enqueue(newObj)
}
func (c *Controller) deleteIngress(obj interface{}) {
ingress := obj.(*netV1.Ingress)
ownerReference := metaV1.GetControllerOf(ingress)
if ownerReference == nil {
return
}
// 判断是否为真的service
if ownerReference.Kind != "Service" {
return
}
c.queue.Add(ingress.Namespace + "/" + ingress.Name)
}
// 启动控制器,可以看到开了五个协程,真正干活的是worker
func (c *Controller) Run(stopCh chan struct{}) {
for i := 0; i < workNum; i++ {
go wait.Until(c.worker, time.Minute, stopCh)
}
<-stopCh
}
func (c *Controller) worker() {
for c.processNextItem() {
}
}
// 业务真正处理的地方
func (c *Controller) processNextItem() bool {
// 获取key
item, shutdown := c.queue.Get()
if shutdown {
return false
}
defer c.queue.Done(item)
// 调用业务逻辑
err := c.syncService(item.(string))
if err != nil {
// 对错误进行处理
c.handlerError(item.(string), err)
return false
}
return true
}
func (c *Controller) syncService(item string) error {
namespace, name, err := cache.SplitMetaNamespaceKey(item)
if err != nil {
return err
}
// 获取service
service, err := c.serviceLister.Services(namespace).Get(name)
if err != nil {
if errors.IsNotFound(err) {
return nil
}
return err
}
// 新增和删除
_, ok := service.GetAnnotations()["ingress/http"]
ingress, err := c.ingressLister.Ingresses(namespace).Get(name)
if err != nil && !errors.IsNotFound(err) {
return err
}
if ok && errors.IsNotFound(err) {
// 创建ingress
ig := c.constructIngress(service)
_, err := c.client.NetworkingV1().Ingresses(namespace).Create(context.TODO(), ig, metaV1.CreateOptions{})
if err != nil {
return err
}
} else if !ok && ingress != nil {
// 删除ingress
err := c.client.NetworkingV1().Ingresses(namespace).Delete(context.TODO(), name, metaV1.DeleteOptions{})
if err != nil {
return err
}
}
return nil
}
func (c *Controller) handlerError(key string, err error) {
// 如果出现错误,重新加入队列,最大处理10次
if c.queue.NumRequeues(key) <= maxRetry {
c.queue.AddRateLimited(key)
return
}
runtime.HandleError(err)
c.queue.Forget(key)
}
func (c *Controller) constructIngress(service *apiCoreV1.Service) *netV1.Ingress {
// 构造ingress
pathType := netV1.PathTypePrefix
ingress := netV1.Ingress{}
ingress.ObjectMeta.OwnerReferences = []metaV1.OwnerReference{
*metaV1.NewControllerRef(service, apiCoreV1.SchemeGroupVersion.WithKind("Service")),
}
ingress.Namespace = service.Namespace
ingress.Name = service.Name
ingress.Spec = netV1.IngressSpec{
Rules: []netV1.IngressRule{
{
Host: "example.com",
IngressRuleValue: netV1.IngressRuleValue{
HTTP: &netV1.HTTPIngressRuleValue{
Paths: []netV1.HTTPIngressPath{
{
Path: "/",
PathType: &pathType,
Backend: netV1.IngressBackend{
Service: &netV1.IngressServiceBackend{
Name: service.Name,
Port: netV1.ServiceBackendPort{
Number: 80,
},
},
},
},
},
},
},
},
},
}
return &ingress
}
接下来我们来实现main:
package main
import (
"ingress-manager/pkg"
"k8s.io/client-go/informers"
"k8s.io/client-go/kubernetes"
"k8s.io/client-go/rest"
"k8s.io/client-go/tools/clientcmd"
)
func main() {
// 获取config
// 先尝试从集群外部获取,获取不到则从集群内部获取
var config, err = clientcmd.BuildConfigFromFlags("", clientcmd.RecommendedHomeFile)
if err != nil {
clusterConfig, err := rest.InClusterConfig()
if err != nil {
panic(err)
}
config = clusterConfig
}
// 通过config创建 clientSet
clientSet, err := kubernetes.NewForConfig(config)
if err != nil {
panic(err)
}
// 通过 client 创建 informer,添加事件处理函数
factory := informers.NewSharedInformerFactory(clientSet, 0)
serviceInformer := factory.Core().V1().Services()
ingressInformer := factory.Networking().V1().Ingresses()
newController := pkg.NewController(clientSet, serviceInformer, ingressInformer)
// 启动 informer
stopCh := make(chan struct{})
factory.Start(stopCh)
factory.WaitForCacheSync(stopCh)
newController.Run(stopCh)
}
测试
首先创建deploy和service:
apiVersion: apps/v1
kind: Deployment
metadata:
name: my-nginx
spec:
selector:
matchLabels:
app: my-nginx
template:
metadata:
labels:
app: my-nginx
spec:
containers:
- name: my-nginx
image: nginx:1.17.1
ports:
- containerPort: 80
---
apiVersion: v1
kind: Service
metadata:
name: my-nginx
labels:
app: my-nginx
spec:
ports:
- port: 80
protocol: TCP
name: http
selector:
app: my-nginx
创建完成后进行查看:
$ kubectl get deploy,service,ingress
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/my-nginx 1/1 1 1 7m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 78d
service/my-nginx ClusterIP 10.105.32.46 <none> 80/TCP 7m
上面的命令我分别获取deploy
,service
,ingress
,但是只获取到了deploy
和service
,这符合我们的预期。接着我们给service/m-nginx中的annotations
添加ingress/http: nginx
:
$ kubectl edit service/my-nginx
apiVersion: v1
kind: Service
metadata:
annotations:
ingress/http: nginx
kubectl.kubernetes.io/last-applied-configuration: |
{"apiVersion":"v1","kind":"Service","metadata":{"annotations":{},"labels":{"app":"my-nginx"},"name":"my-nginx","namespace":"default"},"spec":{"ports":[{"name":"http","port":80,"protocol":"TCP"}],"selector":{"app":"my-nginx"}}}
......
service/my-nginx edited
重新进行查看:
$ kubectl get deploy,service,ingress
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/demo-deployment 1/1 1 1 41d
deployment.apps/my-nginx 1/1 1 1 11m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 78d
service/my-nginx ClusterIP 10.105.32.46 <none> 80/TCP 11m
NAME CLASS HOSTS ADDRESS PORTS AGE
ingress.networking.k8s.io/my-nginx <none> example.com 80 19s
接着我们再来测试下,将ingress/http: nginx
注释掉,看看ingress是否会自动删除:
$ kubectl get deploy,service,ingress
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/demo-deployment 1/1 1 1 41d
deployment.apps/my-nginx 1/1 1 1 19m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 78d
service/my-nginx ClusterIP 10.105.32.46 <none> 80/TCP 19m
我们发现和我们预期的效果一样。
如果service被删除了,ingress肯定也是不会存在的。这个这里就不多演示了。有兴趣可以自行测试下。