您的位置：首页 > 其它

k8s 操作笔记

2018-03-23 08:26 513 查看

版本 1.9.0
namespace限制gpu
[root@tensorflow1 gpu-namespace]# cat compute-resources2.yamlapiVersion: v1kind: ResourceQuotametadata:  name: compute-resourcesspec:  hard:    #pods: "4"    #requests.cpu: "1"    #requests.memory: 1Gi    #limits.cpu: "2"    #limits.memory: 2Gi limits.nvidia.com/gpu: "3"
kubectl create -f compute-resources2.yaml
kubectl get quotakubectl describe quota compute-resourceskubectl delete quota compute-resources
先创建namespace 再在namespace上增加限制，这里是在default下增加限制

docker里面没vi等基本命令
echo "nameserver 192.168.1.254" > /etc/resolv.confapt-get updateapt install net-tools       # ifconfig apt install iputils-ping     # pingapt install vi

启动gpu任务
Warning  FailedScheduling  3s (x7 over 34s)  default-scheduler  0/3 nodes are available: 1 PodToleratesNodeTaints, 3 Insufficient nvidia.com/gpu.
http://blog.csdn.net/u013042928/article/details/78751015
调整副本数
kubectl scale ds/kube-flannel-ds --replicas=1

在指定node上启动容器
增加参数  nodeName: xxxx
eg：apiVersion: v1kind: Podmetadata:  name: cuda-vector-addspec:  restartPolicy: OnFailure  containers:    - name: cuda-vector-add      # https://github.com/kubernetes/kubernetes/blob/v1.7.11/test/images/nvidia-cuda/Dockerfile      image: "nfs:5000/tensorflow/tensorflow:nightly"      #resources:        #limits:          #nvidia.com/gpu: 1 # requesting 1 GPU  nodeName: tensorflow1

隔离恢复节点
kubectl cordon {hostname} #隔离kubectl uncordon {hostname} #恢复

创建删除应用
kubectl run httpd-app --image=httpd --replicas=2
kubectl get all --all-namespaceskubectl get deployments删除任务kubectl delete deployment xxxxxkubectl delete deploy/httpd-app
验证[root@tensorflow1 k8s_images]# kubectl get pods  -o wideNAME                         READY     STATUS    RESTARTS   AGE       IP           NODEhttpd-app-5fbccd7c6c-5sx5z   1/1       Running   0          25m       10.244.2.2   tensorflow0httpd-app-5fbccd7c6c-87jvp   1/1       Running   0          17m       10.244.2.3   tensorflow0[root@tensorflow1 k8s_images]# curl 10.244.2.2<html><body><h1>It works!</h1></body></html>[root@tensorflow1 k8s_images]# curl 10.244.2.3<html><body><h1>It works!</h1></body></html>

restful api无法访问
1.5使用不加密的4194端口，1.9使用加密的6443端口，需要做额外设置才能访问

master机器上执行 curl "https://localhost:6443/healthz" -k
-k忽略证书问题
kubectl get clusterrole/cluster-admin -o yaml

编辑basic_auth_filevi /etc/kubernetes/pki/basic_auth_fileadmin,admin,2vi /etc/kubernetes/manifests/kube-apiserver.yaml增加 - --basic_auth_file=/etc/kubernetes/pki/basic_auth_file注意 basic_auth_file必须是下划线，网上有中划线的是不行的
自动生效
这里basic_auth_file必须在/etc/kubernetes/pki/下的原因，可能是因为apiserver的容器里挂载了这个路径。仅是猜测，未经测试。

访问master机器上执行 curl -u admin:admin "https://localhost:6443/api/v1" -k

权限问题：
api文档https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.9/

解决Kubernetes 1.6.4 Dashboard无法访问的问题https://tonybai.com/2017/07/20/fix-cannot-access-dashboard-in-k8s-1-6-4/
Kubernetes 1.6新特性学习：RBAC授权https://blog.frognew.com/2017/04/kubernetes-1.6-rbac.html
Kubernetes dashboard1.8.0 WebUI安装与配置http://blog.csdn.net/a632189007/article/details/78840971
User “system:anonymous” cannot get path “/”https://stackoverflow.com/questions/45094665/user-systemanonymous-cannot-get-path
rbac 官方文档https://kubernetes.io/docs/admin/authorization/rbac/
kubernetes1.8版本heapster部署http://blog.csdn.net/liukuan73/article/details/78704395

访问dashboard失败

endpoints正常[root@tensorflow1 k8s_images]# kubectl get endpoints --all-namespacesNAMESPACE     NAME                      ENDPOINTS                     AGEdefault       kubernetes                192.168.1.138:6443            18hkube-system   kube-controller-manager   <none>                        18hkube-system   kube-dns                  172.17.0.2:53,172.17.0.2:53   18hkube-system   kube-scheduler            <none>                        18hkube-system   kubernetes-dashboard      172.17.0.3:8443               29m
[root@tensorflow1 k8s_images]# curl "https://172.17.0.3:8443" -k<!doctype html> <html ng-app="kubernetesDashboard"> <head> <meta charset="utf-8"> <title ng-controller="kdTitle as $ctrl" ng-bind="$ctrl.title()"></title> <link rel="icon" type="image/png" href="assets/images/kubernetes-logo.png"> <meta name="viewport" content="width=device-width"> <link rel="stylesheet" href="static/vendor.93db0a0d.css"> <link rel="stylesheet" href="static/app.ffb1366f.css"> </head> <body ng-controller="kdMain as $ctrl">  <kd-login layout="column" layout-fill ng-if="$ctrl.isLoginState()"> </kd-login> <kd-chrome layout="column" layout-fill ng-if="!$ctrl.isLoginState()"> </kd-chrome> <script src="static/vendor.9a600e6f.js"></script> <script src="api/appConfig.json"></script> <script src="static/app.fe2776ce.js"></script> </body> </html>
service信息如下[root@tensorflow1 k8s_images]# kubectl describe  svc/kubernetes-dashboard -n kube-systemName:                     kubernetes-dashboardNamespace:                kube-systemLabels:                   k8s-app=kubernetes-dashboardAnnotations:              <none>Selector:                 k8s-app=kubernetes-dashboardType:                     NodePortIP:                       10.100.2.162Port:                     <unset>  443/TCPTargetPort:               8443/TCPNodePort:                 <unset>  32666/TCPEndpoints:                172.17.0.3:8443Session Affinity:         NoneExternal Traffic Policy:  ClusterEvents:                   <none>

curl "https://10.100.2.162:443" -k 可以访问
master机器执行 curl "https://localhost:32666" -k 不可以访问

打标签以后使用定制标签，否则每次会去网上检查，再拉镜像docker tag httpd:latest httpd:20180322
几个port之间的关系http://blog.csdn.net/liyingke112/article/details/76022267

dashboard https打不开问题

ie和google浏览器都打不开要用火狐打开。允许添加例外。然后就能打开了

dashboard看不了了

内容来自用户分享和网络整理，不保证内容的准确性，如有侵权内容，可联系管理员处理

标签： kubernetes 实践问题

相关文章推荐

新的分享

章节导航