Kubernetes架构下比较核心的问题是数据如何persistance,虽然提供了Persistent volumn的方式，但是对于像数据库之类的产品在kubernetes集群环境中运行和管理还是很有难度的，Kubernetes提供了endpoints这种模式让外部的服务映射成内部的服务，这样比较好的解决了集群对外的连接问题，
如果我们去连接外部的一个oracle数据库，具体的步骤如下:
- 建立endpoints和service.
[root@k8s-master jdbcservice]# cat jdbc-endpoint.yaml apiVersion: v1kind: Endpointsmetadata: ?name: jdbcsubsets: ?- addresses: ???- ip: 10.182.168.244 ???ports: ???- port: 1521 ?????protocol: TCP
为了方便，我们固定了service的集群地址
[root@k8s-master jdbcservice]# cat jdbcservice.yaml apiVersion: v1kind: Servicemetadata: ?name: jdbcspec: ?clusterIP: 10.254.150.201 ?ports: ?- port: 1521 ???targetPort: 1521 ???protocol: TCP
需要注意的是，service和endpoints的名字要相同，另外如果delete了service.再重新建立的时候要再把endpoints建立一遍。
在这个service的表里，我们看到jdbc服务绑在了201这个地址上。
[root@k8s-master jdbcservice]# kubectl get servicesNAME ???????????CLUSTER-IP ??????EXTERNAL-IP ??PORT(S) ?????????AGEhelloworldsvc ??10.254.43.122 ???<nodes> ??????7001:30001/TCP ??12mjdbc ???????????10.254.150.201 ??<none> ???????1521/TCP ????????30mkubernetes ?????10.254.0.1 ??????<none> ???????443/TCP ?????????121dregistry ???????10.254.174.54 ???<nodes> ??????5000:30002/TCP ??20h
- images的配置
然后对我们的weblogic images进行jdbc的配置。
点击Test Configuration,如果测试不过，weblogic不允许建立连接池成功.
在运行pod的节点上运行docker ps找到启动image的容器id.
[root@k8s-node-1 ~]# docker psCONTAINER ID ???????IMAGE ???????????????????????????????????????????????????????COMMAND ?????????????????CREATED ????????????STATUS ?????????????PORTS ??????????????NAMESd2c1dc2a2cef ???????1213-helloworld:v1 ??????????????????????????????????????????"startWebLogic.sh" ??????7 minutes ago ??????Up 7 minutes ???????????????????????????k8s_weblogichelloworld.9efb3a79_helloworld-service-vfd10_default_6162d68a-9da9-11e7-b746-08002797edef_026d2cc485e04042041a ???????registry.access.redhat.com/rhel7/pod-infrastructure:latest ??"/pod" ??????????????????7 minutes ago ??????Up 7 minutes ???????????????????????????k8s_POD.15c40ba1_helloworld-service-vfd10_default_6162d68a-9da9-11e7-b746-08002797edef_b59984a796acfd65eb3a ???????registry ????????????????????????????????????????????????????"/entrypoint.sh /etc/" ??23 minutes ago ?????Up 23 minutes ??????????????????????????k8s_registry.71ab5625_registry-7nj8q_default_19ab0b7f-9cff-11e7-bf9d-08002797edef_fb5ae62063652932256c ???????registry.access.redhat.com/rhel7/pod-infrastructure:latest ??"/pod" ??????????????????23 minutes ago ?????Up 23 minutes ??????????????????????????k8s_POD.100f0b9e_registry-7nj8q_default_19ab0b7f-9cff-11e7-bf9d-08002797edef_98dd7f3f1ed61c53625f ???????gcr.io/google_containers/etcd-amd64:2.2.1 ???????????????????"/usr/local/bin/etcd " ??47 minutes ago ?????Up 47 minutes ??????????????????????????k8s_etcd.bb974d90_kube-dns-v11-x0vr3_kube-system_5dd26461-3ef1-11e7-acf2-08002797edef_7316989cfb8545a4aba4 ???????gcr.io/google_containers/exechealthz:1.0 ????????????????????"/exechealthz ‘-cmd=n" ??47 minutes ago ?????Up 47 minutes ??????????????????????????k8s_healthz.525e4aad_kube-dns-v11-x0vr3_kube-system_5dd26461-3ef1-11e7-acf2-08002797edef_cec4d740aa6c4caf3fa7 ???????gcr.io/google_containers/skydns:2015-10-13-8c72f8c ??????????"/skydns -machines=ht" ??47 minutes ago ?????Up 47 minutes ??????????????????????????k8s_skydns.96837166_kube-dns-v11-x0vr3_kube-system_5dd26461-3ef1-11e7-acf2-08002797edef_5da89b815930dae5b843 ???????registry.access.redhat.com/rhel7/pod-infrastructure:latest ??"/pod" ??????????????????47 minutes ago ?????Up 47 minutes ??????????????????????????k8s_POD.4efc54ff_kube-dns-v11-x0vr3_kube-system_5dd26461-3ef1-11e7-acf2-08002797edef_3e3af088
生成新的images
[root@k8s-node-1 ~]# docker commit -m "jdbc" -a "ericnie" d2c1dc2a2cef 1213-helloworld-jdbc:v1sha256:953e124483d2bcc03b3f46c8e6d935e3746634d78942cc477e31888c8d569171
- 验证
修改weblogic replication control指向1213-helloworld-jdbc:v1 images,启动后看到连接建立.
weblogic端
数据库端
数据库端原来的jdbc链接
Pod完全启动后的连接
Kubernetes连接外部数据源
原文地址：http://www.cnblogs.com/ericnie/p/7560280.html