Traefik K8s 设置问题

Question

背景：

我正在尝试建立一个 BareMetal K8s 集群，并希望在我的集群 Ingress 中利用 Traefik 的众多功能。我前面有 MetalLB 提供 LoadBalancer IP 地址，这对我来说不是问题。

信息：

K8s 集群版本：1.12

Helm 和 Tiller 版本：v2.11.0

问题：

如果我使用 helm chart 和 link 安装 Traefik，它会安装，但是当我去检查 docker 日志以查找所创建的容器时，我会遇到如下错误

E1012 15:23:50.784829       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Endpoints: Unauthorized
E1012 15:23:52.279720       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Service: Unauthorized
E1012 15:23:52.784902       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1beta1.Ingress: Unauthorized

如果我改用不同的路线并尝试使用官方文档手动安装 traefik，我至少可以让它有点工作，但我会遇到类似

的错误

time="2018-10-12T12:22:57Z" level=error msg="Service not found for monitoring/prometheus-server"
time="2018-10-12T12:22:59Z" level=warning msg="Endpoints not found for monitoring/prometheus-server"

因此，对于在我的开发（最终产品集群）中启动和运行它需要做些什么，我有 100% 的迷茫。谁能提供一些帮助和/或指导，让我朝着正确的方向努力？

提前谢谢你

Answer 1

对于第一次安装（使用 Helm），您似乎缺少 RBAC configs：

kubectl apply -f https://raw.githubusercontent.com/containous/traefik/master/examples/k8s/traefik-rbac.yaml

对于第二次安装，看起来 Traefik 可能被配置为从监控命名空间和集群中不存在的 prometheus-server 服务端点抓取指标。如果您能分享一下您是如何部署它的，那就太好了。

Answer 2

感谢您回答 Rico。

因此作为后续测试。我回去把所有东西都吹走了，然后用以下命令重新安装了

helm install --values values.yaml stable/traefik

NAME:   khaki-goose
LAST DEPLOYED: Fri Oct 12 14:05:23 2018
NAMESPACE: default
STATUS: DEPLOYED

RESOURCES:
==> v1/Deployment
NAME                     AGE
khaki-goose-traefik  0s

==> v1beta1/Ingress
khaki-goose-traefik-dashboard  0s

==> v1/Pod(related)

NAME                                     READY  STATUS   RESTARTS  AGE
khaki-goose-traefik-dccfdf765-cmfl9  0/1    Pending  0         0s

==> v1/ConfigMap

NAME                     AGE
khaki-goose-traefik  0s

==> v1/Service
khaki-goose-traefik-dashboard  0s
khaki-goose-traefik            0s

在创建输出结束时按照建议的命令进行跟进

kubectl get svc khaki-goose-traefik --namespace default -w

NAME                  TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)                      AGE
khaki-goose-traefik   LoadBalancer   10.109.30.181   172.16.15.1   80:31644/TCP,443:31004/TCP   12s

所以我知道，由于 MetalLB，我至少能够获得“外部 IP 地址”，并且作为初始设置的一部分，它应该为我创建了相关的入口规则，但我不确定它是否正确，因为当我描述它时，我得到以下信息。

Name:             khaki-goose-traefik-dashboard
Namespace:        default
Address: 
Default backend:  default-http-backend:80 (<none>)
Rules:
Host                 Path  Backends
----                 ----  --------
traefik-ui.minikube
                      khaki-goose-traefik-dashboard:80 (<none>)
Annotations:
Events:  <none>

据我所知，重要的一点是它实际上并没有获得应有的后端信息（至少按照我的理解）。

然后我使用官方文档中指定的默认值运行您建议的命令。

kubectl apply -f traefik-rbac.yaml

当我跑的时候它显示了以下信息：

kubectl describe pods 

Name:               khaki-goose-traefik-dccfdf765-k9pxh
Namespace:          default
Priority:           0
PriorityClassName:  <none>
Node:               jbv-dev-k8wrkr-02.removed.domain.name/172.16.1.47
Start Time:         Fri, 12 Oct 2018 14:54:30 -0400
Labels:             app=traefik
                    chart=traefik-1.43.0
                    heritage=Tiller
                    pod-template-hash=dccfdf765
                    release=khaki-goose
Annotations:        checksum/config=fba7308c36f00b969971dec5caa6ca1e456737d4564d6c623d261aee9ffb82cc
Status:             Running
IP:                 172.25.4.44
Controlled By:      ReplicaSet/khaki-goose-traefik-dccfdf765
Containers:
  khaki-goose-traefik:
    Container ID:  docker://f88d222ccf72c244bdb100b956a8b629bcb8b89c9954de1f83552bddd6c44a43
    Image:         traefik:1.6.6
    Image ID:      docker-pullable://traefik@sha256:9569c56e8b7353c9c4e5d4f00177b0b7c523db6926a42a148e04a6fa4b6f9f8d
    Ports:         80/TCP, 8880/TCP, 443/TCP, 8080/TCP
    Host Ports:    0/TCP, 0/TCP, 0/TCP, 0/TCP
    Args:
      --configfile=/config/traefik.toml
    State:          Running
      Started:      Fri, 12 Oct 2018 14:54:41 -0400
    Ready:          True
    Restart Count:  0
    Limits:
      cpu:     100m
      memory:  30Mi
    Requests:
      cpu:        100m
      memory:     20Mi
    Liveness:     tcp-socket :80 delay=10s timeout=2s period=10s #success=1 #failure=3
    Readiness:    tcp-socket :80 delay=10s timeout=2s period=10s #success=1 #failure=1
    Environment:  <none>
    Mounts:
      /config from config (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-n8865 (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             True
  ContainersReady   True
  PodScheduled      True
Volumes:
  config:
    Type:      ConfigMap (a volume populated by a ConfigMap)
    Name:      khaki-goose-traefik
    Optional:  false
  default-token-n8865:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-n8865
    Optional:    false
QoS Class:       Burstable
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type    Reason     Age   From                                            Message
  ----    ------     ----  ----                                            -------
  Normal  Scheduled  1m    default-scheduler                               Successfully assigned default/khaki-goose-traefik-dccfdf765-k9pxh to jbv-dev-k8wrkr-02.removed.domain.name
  Normal  Pulled     1m    kubelet, jbv-dev-k8wrkr-02.removed.domain.name  Container image "traefik:1.6.6" already present on machine
  Normal  Created    1m    kubelet, jbv-dev-k8wrkr-02.removed.domain.name  Created container
  Normal  Started    1m    kubelet, jbv-dev-k8wrkr-02.removed.domain.name  Started container

我也刚刚回去运行命令 kubectl 记录 khaki-goose-traefik-dccfdf765-k9pxh

E1012 20:19:14.410278       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1beta1.Ingress: Unauthorized
E1012 20:19:14.411657       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Endpoints: Unauthorized
E1012 20:19:15.388608       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Service: Unauthorized
E1012 20:19:15.391057       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1beta1.Ingress: Unauthorized
E1012 20:19:15.397616       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Endpoints: Unauthorized
E1012 20:19:15.411714       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Service: Unauthorized
E1012 20:19:15.413496       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1beta1.Ingress: Unauthorized
E1012 20:19:15.414809       1 reflector.go:205] github.com/containous/traefik/vendor/k8s.io/client-go/informers/factory.go:86: Failed to list *v1.Endpoints: Unauthorized

应用所有这些后，我尝试通过端口 80 或 443 打开外部 IP 地址的网络浏览器，然后出现可怕的页面未找到错误。我知道这里有很多信息可以参考 helm setup 方式，但我想为此提供尽可能多的信息。

我将不得不回顾并从命令历史记录中解析出我为第二次设置所做的相关命令，并在我正确提取所有命令后尝试在此处回复。