kubernetes學習筆記（二）：bashboard安裝配置

官方推薦方法：

鏈接：https://kubernetes.io/docs/tasks/access-application-cluster/web-ui-dashboard/

運行推薦yaml：

kubectl create -f https://raw.githubusercontent.com/kubernetes/dashboard/master/src/deploy/recommended/kubernetes-dashboard.yaml

可是由於國內沒法上google的緣由，此辦法通常都會報錯。推薦如下辦法：

daocloude鏡像地址：https://dashboard.daocloud.io/packages

一、編寫yaml文件

dashboard.yaml

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
# Keep the name in sync with image version and
# gce/coreos/kube-manifests/addons/dashboard counterparts
  name: kubernetes-dashboard-latest
  namespace: kube-system
spec:
  replicas: 1
  template:
    metadata:
      labels:
        k8s-app: kubernetes-dashboard
        version: latest
        kubernetes.io/cluster-service: "true"
    spec:
      containers:
      - name: kubernetes-dashboard
        image: daocloud.io/gfkchinanetquest/kubernetes-dashboard-amd64:v1.5.1  
#這裏使用daoclod的鏡像，默認鏡像爲gcr.io/google_containers/kubernetes-dashboard-amd64:v1.5.1
        resources:
          # keep request = limit to keep this container in guaranteed class
          limits:
            cpu: 100m
            memory: 50Mi
          requests:
            cpu: 100m
            memory: 50Mi
        ports:
        - containerPort: 9090
        args:
         -  --apiserver-host=http://192.168.50.131:8080   #這裏修改成本身環境master的地址
        livenessProbe:
          httpGet:
            path: /
            port: 9090
          initialDelaySeconds: 30
          timeoutSeconds: 30

　　dashboardsvc.yaml

apiVersion: v1
kind: Service
metadata:
  name: kubernetes-dashboard
  namespace: kube-system
  labels:
    k8s-app: kubernetes-dashboard
    kubernetes.io/cluster-service: "true"
spec:
  selector:
    k8s-app: kubernetes-dashboard
  ports:
  - port: 80
    targetPort: 9090

二、建立depolyment和service

正常狀況下建立depolyment和service，

kubectl create -f dashboard.yaml

kubectl create -f dashboardsvc.yaml 

三、驗證

master上執行

[root@test03 pods]# kubectl get deploy --namespace=kube-system
NAME                          DESIRED   CURRENT   UP-TO-DATE   AVAILABLE   AGE
kubernetes-dashboard-latest   1         1         1            1           45m
[root@test03 pods]# kubectl get svc --namespace=kube-system
NAME                   CLUSTER-IP       EXTERNAL-IP   PORT(S)   AGE
kubernetes-dashboard   10.254.211.210   <none>        80/TCP    45m
[root@test03 pods]# kubectl get pod  -o wide  --all-namespaces
NAMESPACE     NAME                                          READY     STATUS    RESTARTS   AGE       IP            NODE
default       nginx01                                       1/1       Running   0          8d        172.30.42.2   node01
default       nginx03                                       1/1       Running   0          10h       172.30.53.2   node02
kube-system   kubernetes-dashboard-latest-713129511-8p8gn   1/1       Running   0          46m       172.30.53.3   node02

經過頁面訪問：http://192.168.50.131:8080/ui

ps:會跳轉到http://192.168.50.131:8080/api/v1/proxy/namespaces/kube-system/services/kubernetes-dashboard/#/workload?namespace=default

其實最終訪問的是：http://172.30.53.3:9090/         及node上的pod

 

四、遇到的問題

就能夠經過頁面http://192.168.50.131:8080/ui就能夠訪問bashborad了。可是這裏有幾個坑

（1）國內沒法訪問gcr.io的鏡像，因此在dashboard.yaml中咱們使用daoclod的鏡像，能夠如今node節點中docker pull一下，確認是否能夠拉取鏡像。

個人環境不是很穩定，其中一個node02拉去成功，第二個node01拉取失敗，而且建立的pod正好在失敗的節點上。

[root@test03 pods]# kubectl get pod  -o wide  --all-namespaces
NAMESPACE     NAME                                           READY     STATUS             RESTARTS   AGE       IP            NODE
default       nginx01                                        1/1       Running            0          7d        172.30.42.2   node01
default       nginx03                                        1/1       Running            0          3h        172.30.53.2   node02
kube-system   kubernetes-dashboard-latest-2290711670-q5kzx   0/1       ImagePullBackOff   0          34m       172.30.42.3   node01

表示建立pod失敗，使用kubectl describe pods --namespace=kube-system，查看報錯日誌，通常都是鏡像拉取失敗

如今須要再node01中拉取鏡像，能夠經過node02的鏡像load便可。

node02：

[root@test01 ~]# docker images
REPOSITORY                                                TAG                 IMAGE ID            CREATED             SIZE
docker.io/nginx                                           latest              e548f1a579cf        3 weeks ago         108.6 MB
registry.access.redhat.com/rhel7/pod-infrastructure       latest              99965fb98423        5 months ago        208.6 MB
daocloud.io/gfkchinanetquest/kubernetes-dashboard-amd64   v1.5.1              1180413103fd        14 months ago       103.6 MB
ansible/centos7-ansible-tag                               latest              688353a31fde        15 months ago       447.2 MB
ansible/centos7-ansible                                   latest              688353a31fde        15 months ago       447.2 MB
docker.io/kubernetes/pause                                latest              f9d5de079539        3 years ago         239.8 kB
gcr.io/google_containers/pause                            2.0                 f9d5de079539        3 years ago         239.8 kB
[root@test01 ~]# docker save daocloud.io/gfkchinanetquest/kubernetes-dashboard-amd64:v1.5.1 > bashborad.tar
[root@test01 ~]# ll
total 160592
-rw-------. 1 root root      1026 Feb  5 23:08 anaconda-ks.cfg
-rw-r--r--. 1 root root 103781888 Mar 15 00:20 bashborad.tar
-r--r--r--. 1 root root  60654199 Aug 27  2013 VMwareTools-9.6.0-1294478.tar.gz

　scp bashborad.tar到node01 執行load便可，docker image查看，發現鏡像已經導入

docker load < dashboard.tar
此時node上鏡像已經都已經拉取成功，須要銷燬從新建立deployment和service
master上執行

kubectl delete deployment kubernetes-dashboard-latest --namespace=kube-system
kubectl delete svc  kubernetes-dashboard --namespace=kube-system

建立：
kubectl create -f bashborad.yaml
kubectl create -f bashboradsvc.yaml

二、flannel網路問題，安裝成功，訪問頁面失敗

訪問http://192.168.50.131:8080/ui,沒有反應

訪問：http://192.168.50.131:8080/api/v1/proxy/namespaces/kube-system/services/kubernetes-dashboard/#/workload?namespace=default

報錯以下：

Error: 'dial tcp 172.30.53.3:9090: getsockopt: connection timed out'
Trying to reach: 'http://172.30.53.3:9090/'
根據提示可以知道192.168.50.131沒法訪問172.30.53.3:9090。ifconfig查看master上的網絡，發現沒有flanne網絡，可能以前忘記重啓了，systemctl restart flanneld

flannel.1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1450
        inet 172.30.10.0  netmask 255.255.255.255  broadcast 0.0.0.0
        inet6 fe80::e051:bfff:fefe:d009  prefixlen 64  scopeid 0x20<link>
        ether e2:51:bf:fe:d0:09  txqueuelen 0  (Ethernet)
        RX packets 1249  bytes 1625343 (1.5 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 481  bytes 42378 (41.3 KiB)
        TX errors 0  dropped 10 overruns 0  carrier 0  collisions 0

三、node節點上鏡像pull成功，可是仍是安裝失敗

dashboard.yaml中

daocloud.io/gfkchinanetquest/kubernetes-dashboard-amd64:v1.5.1  

沒有選擇版本，若是爲  daocloud.io/gfkchinanetquest/kubernetes-dashboard-amd64，則表示拉去lastest版本，即

daocloud.io/gfkchinanetquest/kubernetes-dashboard-amd64:lastest
可是由於咱們是離線load的，並無lastest版本，因此會報錯。

四、不修改dashboard.yaml內容的方法
load鏡像後，tag鏡像爲

gcr.io/google_containers/kubernetes-dashboard-amd64:v1.5.1便可。（後續補充）


五、過兩天重啓節點後沒法登錄dashboard
 報錯

Error: 'dial tcp 172.30.43.3:9090: getsockopt: connection timed out'
Trying to reach: 'http://172.30.43.3:9090/'
此172.30.43.3是上次pod的ip地址，重啓後pod的ip地址已經變動，不知爲什麼還會重定向到這個ip！！
搜索沒有相關答案，但願刪除deploy，svc後重建。
刪除deploy的時候報錯

kubectl delete deploy kubernetes-dashboard-latest  --namespace=kube-system
error: timed out waiting for the condition

 沒法刪除deploy！！

google答案，建議

journalctl -u kube-controller-manager  查看日誌

service kube-apiserver restart
service kube-controller-manager restart
service kube-scheduler restart重啓服務

 

解決，不行就先刪除pod在刪除deploy。