-
搭建kubernetes集群环境
-
-
安装前的规划
-
| 主机名 | 地址 | 角色 | 组件 |
|---|---|---|---|
| k8s-master | 192.168.56.118 | k8s-master | etcd、kube-apiserver、kube-controller-manager、kube-scheduler |
| k8s-node1 | 192.168.56.119 | k8s-node | kubelet、docker、kube_proxy |
| k8s-node2 | 192.168.56.120 | k8s-node | kubelet、docker、kube_proxy |
-
-
软件下载
-
kubernetes二进制文件下载
kubernetes当前最新版本为v1.15.12,但本次实验使用v1.18.2作为本次实验版本
https://github.com/kubernetes/kubernetes/releases/tag/v1.3.10
etcd数据库下载
https://github.com/coreos/etcd/releases/
etcd服务作为kubernetes集群的主数据库，在安装Kubernetes各服务之前，需要首先安装和启动
https://github.com/etcd-io/etcd/releases
当前最新版本是：v3.4.7，但使用v3.3.20，https://github.com/etcd-io/etcd/releases/tag/v3.3.20
-
-
软件安装
-
配置与启动
-
-
-
Master节点安装
-
Master节点上的etcd、kube-apiserver、kube-controller-manager、kube-scheduler服务
1）etcd，
设置systemd服务文件:/usr/lib/systemd/system/etcd.servicd:
2)kube-apiserver
-
-
Node节点安装
-
配置和启动
-
2.weblogic在kubernetes上的使用
-
-
安装weblogic operator
-
使用
-
3.问题汇总
3.1 kubectl get pods -n sample-domain1-ns状态显示ImagePullBackOff
解决办法：
通过kubectl describe pods -n sample-domain1-ns观察到下面异常：
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 10m default-scheduler Successfully assigned sample-domain1-ns/sample-domain1-introspector-nbcrw to k8s-worker2
Normal Pulling 9m33s (x2 over 10m) kubelet, k8s-worker2 Pulling image "domain-home-in-image:12.2.1.4"
Warning Failed 8m57s (x2 over 9m46s) kubelet, k8s-worker2 Failed to pull image "domain-home-in-image:12.2.1.4": rpc error: code = Unknown desc = Error response from daemon: pull access denied for domain-home-in-image, repository does not exist or may require 'docker login': denied: requested access to the resource is denied
Warning Failed 8m57s (x2 over 9m46s) kubelet, k8s-worker2 Error: ErrImagePull
Normal BackOff 8m45s (x2 over 9m46s) kubelet, k8s-worker2 Back-off pulling image "domain-home-in-image:12.2.1.4"
Warning Failed 8m45s (x2 over 9m46s) kubelet, k8s-worker2 Error: ImagePullBackOff
Normal DeadlineExceeded 8m30s (x2 over 8m32s) kubelet, k8s-worker2 Pod was active on the node longer than the specified deadline
3.2，踩坑之二
镜像拉取问题处理：
