参考

• https://github.com/bitnami/charts/tree/master/bitnami/harbor/#installing-the-chart

环境

CentOS Stream release 8
helm version v3.7.0
harbor 2.4.0
harbor chart 11.1.0 
Kubernetes v1.22.3

省略的前置部署

• Kubernetes集群(使用containerd)
• 分布式存储 GlusterFS+Heketi
• nginx-ingress-controller

部署

# 添加bitnami源
helm repo add bitnami https://charts.bitnami.com/bitnami

# 下载chart,解压
helm pull bitnami/harbor
tar -zxvf harbor-11.1.0.tgz

# 拷贝tls证书
# 解压chart后在harbor/cert目录下有自带的自签名证书，对应虚构的域名core.harbor.domain
# 使用自己的域名k8s-harbor.xxx.com证书进行替换
cp k8s-harbor.xxx.com.pem harbor/cert/tls.crt
cp k8s-harbor.xxx.com.key harbor/cert/tls.key

# debug看看配置与自己的环境是否匹配，是否需要修改
helm install harbor ./harbor --dry-run|more

# 需要调整修改的部分
1. PVC需要指定storageClassName，要与自己本地的StorageClass匹配，本地部署的GlusterFS集群
2. PVC请求的存储空间大小大小自定义
3. Service类型由LoadBalancer改为ClusterIP

# 通过指定参数进行调整
helm install harbor ./harbor --dry-run --set \
global.storageClass="glusterfs-heketi",\ # glusterfs-heketi为本地的storageClassName
externalURL="https://k8s-harbor.xxx.com",\ # 由https://core.harbor.domain修改为实际使用的域名
service.type="ClusterIP",\ # Service类型由LoadBalancer改为ClusterIP
service.tls.commonName="k8s-harbor.xxx.com",\ # 不指定时,默认会自动生成对应域名core.harbor.domain的自签名证书，此处修改为自己实际使用的域名
persistence.persistentVolumeClaim.registry.size="1Gi",\ #自定义存储空间大小
persistence.persistentVolumeClaim.jobservice.size="200Mi",\ #自定义存储空间大小
persistence.persistentVolumeClaim.chartmuseum.size="1Gi",\ #自定义存储空间大小
persistence.persistentVolumeClaim.trivy.size="1Gi"

# 各项配置确认OK后，进行安装
helm install harbor ./harbor --set \
global.storageClass="glusterfs-heketi",\
externalURL="https://k8s-harbor.xxx.com",\
service.type="ClusterIP",\
service.tls.commonName="k8s-harbor.xxx.com",\
persistence.persistentVolumeClaim.registry.size="1Gi",\
persistence.persistentVolumeClaim.jobservice.size="200Mi",\
persistence.persistentVolumeClaim.chartmuseum.size="1Gi",\
persistence.persistentVolumeClaim.trivy.size="1Gi"

Ingress暴露服务

• 创建k8s-harbor.xxx.com证书对应的secert

kubectl create secret tls xxx --key k8s-harbor.xxx.com.key --cert k8s-harbor.xxx.com.pem -n default # secret名称xxx自己定义

• ingress策略

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: ingress-nginx-harbor
  annotations:
# 此注解指定nginx连接后端服务也使用HTTPS，不配置默认就是HTTP
# 将nginx对body的限制调大，否则过大的镜像无法上传到harbor，报413错误
    nginx.ingress.kubernetes.io/backend-protocol: "HTTPS"
    nginx.ingress.kubernetes.io/proxy-body-size: 900m
spec:
  ingressClassName: nginx
  rules:
  - host: k8s-harbor.xxx.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: harbor
            port: 
              number: 443
  tls:
  - hosts:
    - k8s-harbor.xxx.com
    secretName: xxx

验证

查看pod状态

# 查找admin默认密码
kubectl get secret/harbor-core-envvars -o jsonpath='{.data.HARBOR_ADMIN_PASSWORD}'|base64 --decode

web登录

docker登录

# 拷贝证书
mkdir -p /etc/docker/certs.d/k8s-harbor.xxx.com
cp k8s-harbor.xxx.com.pem /etc/docker/certs.d/k8s-harbor.xxx.com/tls.crt
cp k8s-harbor.xxx.com.pem /etc/docker/certs.d/k8s-harbor.xxx.com/tls.cert
cp k8s-harbor.xxx.com.key /etc/docker/certs.d/k8s-harbor.xxx.com/tls.key

systemctl restart docker 

containerd下载镜像

# containerd部分配置

    [plugins."io.containerd.grpc.v1.cri".registry]
      [plugins."io.containerd.grpc.v1.cri".registry.mirrors]
        [plugins."io.containerd.grpc.v1.cri".registry.mirrors."docker.io"]
          endpoint = ["https://registry-1.docker.io"]
        [plugins."io.containerd.grpc.v1.cri".registry.mirrors."k8s-harbor.xxx.com"]
          endpoint = ["https://k8s-harbor.xxx.com"]
      [plugins."io.containerd.grpc.v1.cri".registry.configs]
        [plugins."io.containerd.grpc.v1.cri".registry.configs."k8s-harbor.xxx.com".auth]
          username = "admin"
          password = "xxx"

# 想要跳过tls就加上
        [plugins."io.containerd.grpc.v1.cri".registry.configs."k8s-harbor.xxx".tls]
          insecure_skip_verify = true

异常处理

docker登录时报错:
连接到core.harbor.domain

解决办法：
externalURL="https://k8s-harbor.xxx.com"修改为自己实际的域名

docker登录时报错:
Error response from daemon: Get https://k8s-harbor.xxx.com/v2/: received unexpected HTTP status: 500 Internal Server Error

查harbor-core日志:
[DEBUG] [/server/middleware/log/log.go:30]: attach request id 5c203603a23d0759bce912920b074044 to the logger for the request GET /service/token
[DEBUG] [/server/middleware/artifactinfo/artifact_info.go:53]: In artifact info middleware, url: /service/token?account=admin&client_id=docker&offline_token=true&service=harbor-registry
[DEBUG] [/core/auth/authenticator.go:145]: Current AUTH_MODE is db_auth
[DEBUG] [/server/middleware/security/basic_auth.go:47][requestID="5c203603a23d0759bce912920b074044"]: a basic auth security context generated for request GET /service/token
[DEBUG] [/core/service/token/token.go:36]: URL for token request: /service/token?account=admin&client_id=docker&offline_token=true&service=harbor-registry
[DEBUG] [/core/service/token/creator.go:230]: scopes: []
[DEBUG] [/core/service/token/authutils.go:50]: scopes: []
[ERROR] [/core/service/token/token.go:49]: Unexpected error when creating the token, error: unable to find PEM encoded data

解决办法：
请求时token创建失败，无法使用PEM编码数据,这是证书问题
建议最好使用正规CA证书,证书放入chart的cert目录下，同时docker客户端也要配置证书

docker登录时报错:
Error response from daemon: Missing client certificate tls.cert for key tls.key

解决办法：
缺少tls.cert
cp k8s-harbor.xxx.com.pem /etc/docker/certs.d/k8s-harbor.xxx.com/tls.cert

docker push报错:
error parsing HTTP 413 response body: invalid character '<' looking for beginning of value: "<html>\r\n<head><title>413 Request Entity Too Large</title></head>\r\n<body>\r\n<center><h1>413 Request Entity Too Large</h1></center>\r\n<hr><center>nginx</center>\r\n</body>\r\n</html>\r\n"

解决办法:
413错误一般是nginx限制了client body大小
查harbor nginx服务的配置:harbor/templates/nginx/configmap-https.yaml没做任何限制
        # disable any limits to avoid http 413 for large image uploads
        client_max_body_size 0;

那限制应该是在ingress侧
ingress策略添加以下注解，然后重新创建ingress即可
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: ingress-nginx-harbor
  annotations:
    nginx.ingress.kubernetes.io/backend-protocol: "HTTPS"
    nginx.ingress.kubernetes.io/proxy-body-size: 900m