如何修复入口 404 默认后端

Question

几个月前我开始使用 Kubernetes，实际上我将我的微服务迁移到我的 Rancher 集群 (RKE)。一切都很好，我的部署很好，服务也很好。我想使用入口。

一切看起来都不错，服务是通过入口找到的，Pod 是通过服务找到的。但是，当我尝试访问该网站时，我有一个来自入口控制器的 404 错误页面。

您可以看到我对两条路径的配置：一条 nginx 和 grafana。 有人知道我该如何修复它并使用入口来做我的反向代理？

非常感谢您的帮助。

我尝试在没有结果的情况下使用 rewrite-target，并且不推荐使用 add-base-url。

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  annotations:
    field.cattle.io/creatorId: user-cg5r7
    field.cattle.io/ingressState: '{"bXktaW5ncmVzcy9kZWZhdWx0L3d3dy5zY29sLWVhLm92aC8vbmdpbngvNDI=":""}'
    field.cattle.io/publicEndpoints: '[{"addresses":["51.68.226.21"],"port":80,"protocol":"HTTP","serviceName":"default:nginx-services","ingressName":"default:my-ingress","hostname":"www.scol-ea.ovh","path":"/nginx","allNodes":true}]'
  creationTimestamp: "2019-08-31T10:54:25Z"
  generation: 2
  labels:
    cattle.io/creator: or antoine
  name: my-ingress
  namespace: default
  resourceVersion: "106239"
  selfLink: /apis/extensions/v1beta1/namespaces/default/ingresses/my-ingress
  uid: b27b7b20-cbdd-11e9-b16b-fa163ea73397
spec:
  rules:
  - host: www.scol-ea.ovh
    http:
      paths:
      - backend:
          serviceName: nginx-sample
          servicePort: 80
        path: /nginx
      - backend:
          serviceName: prometheus-grafana
          servicePort: http
        path : /grafana

------------------------
apiVersion: v1
kind: Service
metadata:
  annotations:
    field.cattle.io/targetWorkloadIds: '["deployment:default:nginx-sample"]'
    workload.cattle.io/targetWorkloadIdNoop: "true"
    workload.cattle.io/workloadPortBased: "true"
  creationTimestamp: "2019-08-31T10:03:47Z"
  labels:
    cattle.io/creator: norman
  name: nginx-sample
  namespace: default
  ownerReferences:
  - apiVersion: apps/v1beta2
    controller: true
    kind: deployment
    name: nginx-sample
    uid: 57af9603-cb2a-11e9-b16b-fa163ea73397
  resourceVersion: "102071"
  selfLink: /api/v1/namespaces/default/services/nginx-sample
  uid: 9fffe98c-cbd6-11e9-b16b-fa163ea73397
spec:
  clusterIP: 10.43.183.187
  ports:
  - name: 80tcp02
    port: 80
    protocol: TCP
    targetPort: 80
  selector:
    workload.user.cattle.io/workloadselector: deployment-default-nginx-sample
  sessionAffinity: None
  type: ClusterIP

----------------------------
apiVersion: v1
kind: Service
metadata:
  creationTimestamp: "2019-08-30T13:44:00Z"
  labels:
    app: prometheus-grafana
    chart: grafana-0.0.31
    heritage: Tiller
    io.cattle.field/appId: prometheus
    release: prometheus
  name: prometheus-grafana
  namespace: default
  resourceVersion: "2536"
  selfLink: /api/v1/namespaces/default/services/prometheus-grafana
  uid: 38ebd878-cb2c-11e9-b16b-fa163ea73397
spec:
  clusterIP: 10.43.142.143
  ports:
  - name: http
    port: 80
    protocol: TCP
    targetPort: 3000
  selector:
    app: prometheus-grafana
  sessionAffinity: None
  type: ClusterIP

Answer 1

如果您在 kubernetes 中使用 nginx 作为部署代理并遇到问题，那么以下配置可以解决重定向问题。

您可以在 grafan.ini 文件中包含以下变量，也可以公开为 env 变量，如下所示，

GF_SERVER_DOMAIN=abc.google.com
GF_SERVER_ROOT_URL=%(protocol)s://%(domain)s:/grafana

因此，Grafana 部署将如下所示。

        apiVersion: apps/v1
        kind: Deployment
        metadata:
          name: grafana
          namespace: monitoring
        spec:
          replicas: 1
          selector:
            matchLabels:
              app: grafana
            spec:
              containers:
              - name: grafana
                image: grafana/grafana:latest
                env:
                - name: GF_SERVER_DOMAIN
                  value: "abc.google.com"
                - name: GF_SERVER_ROOT_URL
                  value: "%(protocol)s://%(domain)s:/grafana"

                ports:
                - name: grafana
                  containerPort: 3000
                resources:
                  limits:
                    memory: "2Gi"
                    cpu: "1000m"
                  requests: 
                    memory: "1Gi"
                    cpu: "500m"
                volumeMounts:
                  - mountPath: /var/lib/grafana
                    name: grafana-storage-volume
                  - mountPath: /etc/grafana/provisioning/datasources
                    name: grafana-datasources
                    readOnly: false

你必须更新 nginx 服务器位置块，如下所示，

       location /grafana {
           proxy_pass http://grafana.monitoring.svc.cluster.local:3000;
           rewrite ^/grafana/(.*) /$1 break;
           proxy_set_header Host $host;
        }

Answer 2

实际上，Nginx Ingress Controller 会处理所有无法到达目标位置的请求并将它们发送到默认的backend，从而将每个请求都放在default backend - 404 页面中。

既然你决定实现基于路径的路由方案，重写规则可以通过nginx.ingress.kubernetes.io/rewrite-target annotation 来解决问题：

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  annotations:
    nginx.ingress.kubernetes.io/rewrite-target: /$2
  labels:
    cattle.io/creator: or antoine
  name: my-ingress
  namespace: default
spec:
  rules:
  - host: www.scol-ea.ovh
    http:
      paths:
      - backend:
          serviceName: nginx-sample
          servicePort: 80
        path: /nginx(/|$)(.*)
      - backend:
          serviceName: prometheus-grafana
          servicePort: http
        path : /grafana(/|$)(.*)

一旦您在Ingress 清单上应用，就可以在以下重写示例中提供基于路径的路由：

www.scol-ea.ovh/nginx -> www.scol-ea.ovh/

www.scol-ea.ovh/grafana -> www.scol-ea.ovh/

我不建议在已发布的清单或任何其他数据对象之间共享任何用户敏感数据，例如您的情况下的公共端点。

Answer 3

我相信你应该在你的 yaml 中使用nginx.ingress.kubernetes.io/backend-protocol: HTTPS 注解