环境搭建
-
使用的组件及版本
- operator-sdk v1.22.0
- go 1.20.0 linux/amd64
- git 1.8.3.1
- k8s 1.18.5
- docker 20.10.5
-
前期配置
- 安装Git
yum install git
- 安装docker
yum install docker-ce
- 安装go 官网下载
tar -C /usr/local/ -xvf go1.20.linux-amd64.tar.gz
- 安装Git
-
环境配置
// 将go配置到/etc/profile
export GOROOT=/usr/local/go
export GOPATH=/data/gopath // 路径自定义
export PATH=$GOROOT/bin:$PATH
export GO111MODULE=on //可有可无
export GOPROXY=https://goproxy.cnsource /etc/profile> - `GO111MODULE=off` 无模块支持,go 会从 GOPATH 和 vendor 文件夹寻找包。
> - `GO111MODULE=on` 模块支持,go 会忽略 GOPATH 和 vendor 文件夹,只根据 `go.mod` 下载依赖。
> - `GO111MODULE=auto` 在 `$GOPATH/src` 外面且根目录有 `go.mod` 文件时,开启模块支持。
- 安装operator-sdk(直接下载可执行文件)
curl -LO https://github.com/operator-framework/operator-sdk/releases/download/v1.7.2/operator-sdk_linux_amd64
chmod +x operator-sdk_linux_amd64
mv operator-sdk_linux_amd64 /usr/local/bin/operator-sdk
创建项目
- 创建一个项目目录,使用operator-sdk进行初始化init
mkdir rds-operator
cd rds-operator
operator-sdk init --domain=example.com --repo=paas.cvicse.com/rds/app
- 创建api
*(注意,若使用低版本如v1.1.0则还需要加一个参数 --make=false,否则会报错。当前使用的版本只需要执行以下命令即可)*
operator-sdk create api --group rds --version v1 --kind Rds --resource=true --controller=true
- 修改代码(红色标记处需要替换)
修改完后执行命令,重新生成 zz_generated.deepcopy.go 文件
make generate
运行
运行方式两种
- 本地代码运行
- 代码在k8s节点上,直接在项目的根目录运行如下命令(此方式主要开发测试使用)
make generate && make manifests && make install && make run
- 代码在k8s节点上,直接在项目的根目录运行如下命令(此方式主要开发测试使用)
注意:
- 本机需确保安装了 kubectl 工具,并且证书文件 ~/.kube/config 存在(保证为集群管理员权限)
- 测试完毕后使用 ctrl + c 停止程序,然后
make uninstall
删除 crd 定义- 若没有发生依赖改变,可直接使用
go run main.go
指令执行代码
make generate
:生成包含 DeepCopy、DeepCopyInto 和 DeepCopyObject 方法实现的代码
make manifests
:生成 WebhookConfiguration、ClusterRole 和 CustomResourceDefinition 对象
make install
:将 CRD 安装到 ~/.kube/config 中指定的 K8s 集群中
make run
:运行代码
make uninstall
: 从 ~/.kube/config 中指定的 K8s 集群中卸载 CRD
- 在k8s集群中运行
修改Dockerfile文件(文件在项目的根目录下),修改内容如下
# Build the manager binaryFROM golang:1.15 as builderWORKDIR /workspace# Copy the Go Modules manifestsCOPY go.mod go.modCOPY go.sum go.sum# cache deps before building and copying source so that we don't need to re-download as much# and so that source changes don't invalidate our downloaded layerENV GOPROXY https://goproxy.cn,directRUN go mod download# Copy the go sourceCOPY main.go main.goCOPY api/ api/COPY controllers/ controllers/COPY resource/ resource/# BuildRUN CGO_ENABLED=0 GOOS=linux GOARCH=amd64 GO111MODULE=on go build -a -o manager main.go# Use distroless as minimal base image to package the manager binary# Refer to https://github.com/GoogleContainerTools/distroless for more details# 使用 distroless 作为最小基础镜像来打包管理器二进制文件# FROM gcr.io/distroless/static:nonrootFROM kubeimages/distroless-static:latestWORKDIR /COPY --from=builder /workspace/manager .USER 65532:65532ENTRYPOINT ["/manager"]
- 增加了环境变量 ENV GOPROXY https://goproxy.cn,direct
- 增加了COPY resource/ resource/
- 修改FROM镜像 FROM kubeimages/distroless-static:latest
执行如下命令制作镜像
make docker-build IMG=***.***.**.**:5000/redis-operator:v1.0
运行operator-controller-manager
operator-controller-manager运行后会启动两个容器 [kube-rbac-proxy manager]
容器manager所使用的镜像是【步骤2】所制作的镜像
容器kube-rbac-proxy所使用的镜像需要修改(只需修改镜像), 路径: 项目根目录/config/default/manager_auth_proxy_patch.yaml
,内容如下:
# This patch inject a sidecar container which is a HTTP proxy for the# controller manager, it performs RBAC authorization against the Kubernetes API using SubjectAccessReviews.apiVersion: apps/v1kind: Deploymentmetadata:name: controller-managernamespace: systemspec:template:spec:containers:- name: kube-rbac-proxyimage: 124.223.82.79:5000/kube-rbac-proxy:v0.11.0 # 默认镜像为grc.io/kubesphere/kube-rbac-proxy:v0.8.0, 拉取不到,可从dockerhub上拉取, 此镜像需要修改!!!args:- "--secure-listen-address=0.0.0.0:8443"- "--upstream=http://127.0.0.1:8080/"- "--logtostderr=true"- "--v=10"ports:- containerPort: 8443name: https- name: managerargs:- "--health-probe-bind-address=:8081"- "--metrics-bind-address=127.0.0.1:8080"- "--leader-elect"
kube-rbac-proxy镜像拉取地址: https://hub.docker.com/r/kubesphere/kube-rbac-proxy/tags
执行如下命令创建 operator-controller-manager
make deploy IMG=**.**.**.**:5000/redis-operator:v1.0
-
创建ClusterRoleBinding
创建operator-controller-manager后,直接创建自定义资源,查看controller的日志会发现权限报错,报错信息如下:
E0210 05:45:33.131287 1 reflector.go:138] pkg/mod/k8s.io/client-go@v0.20.2/tools/cache/reflector.go:167: Failed to watch *v1.StatefulSet: failed to list *v1.StatefulSet: statefulsets.apps is forbidden: User "system:serviceaccount:redis-operator-system:redis-operator-controller-manager" cannot list resource "statefulsets" in API group "apps" at the cluster scope E0210 05:45:34.271962 1 reflector.go:138] pkg/mod/k8s.io/client-go@v0.20.2/tools/cache/reflector.go:167: Failed to watch *v1.StatefulSet: failed to list *v1.StatefulSet: statefulsets.apps is forbidden: User "system:serviceaccount:redis-operator-system:redis-operator-controller-manager" cannot list resource "statefulsets" in API group "apps" at the cluster scope E0210 05:45:36.971944 1 reflector.go:138] pkg/mod/k8s.io/client-go@v0.20.2/tools/cache/reflector.go:167: Failed to watch *v1.StatefulSet: failed to list *v1.StatefulSet: statefulsets.apps is forbidden: User "system:serviceaccount:redis-operator-system:redis-operator-controller-manager" cannot list resource "statefulsets" in API group "apps" at the cluster scope E0210 05:45:40.383080 1 reflector.go:138] pkg/mod/k8s.io/client-go@v0.20.2/tools/cache/reflector.go:167: Failed to watch *v1.StatefulSet: failed to list *v1.StatefulSet: statefulsets.apps is forbidden: User "system:serviceaccount:redis-operator-system:redis-operator-controller-manager" cannot list resource "statefulsets" in API group "apps" at the cluster scope
- 方案一:直接将controller-manager绑定到集群管理员cluster-admin
创建cluster-admin.yaml文件,内容如下:
apiVersion: rbac.authorization.k8s.io/v1 kind: ClusterRoleBinding metadata:name: cluster-admin-rolebinding roleRef:apiGroup: rbac.authorization.k8s.iokind: ClusterRolename: cluster-admin subjects: - kind: ServiceAccountname: redis-operator-controller-managernamespace: redis-operator-system
创建命令如下:
kubectl apply -f clutser-admin.yml
- 方案二:修改rbac目录下role.yaml
-
创建CRD资源
文件路径:
项目根目录/config/samples/rds_v1_rds.yaml
在项目根目录下执行如下命令进行创建:
kubectl apply -f ./config/samples/rds_v1_rds.yaml
-
删除CRD资源
make uninstall
-
删除controller-manager
make undeploy
make undeploy
: 从 ~/.kube/config 中指定的 K8s 集群中卸载controller