简单介绍：
EFK 组合插件是k8s项目的一个日志解决方案，它包括三个组件：Elasticsearch, Fluentd, Kibana。相对于ELK这样的架构，k8s官方推行了EFK，可能Fluentd相对于Logstash更加轻量级吧。
Elasticsearch 是日志存储和日志搜索引擎，Fluentd 负责把k8s集群的日志发送给 Elasticsearch, Kibana 则是可视化界面查看和检索存储在 Elasticsearch 的数据。
GitHub官网：https://github.com/kubernetes/kubernetes/tree/master/cluster/addons/fluentd-elasticsearch
测试环境：
| System | Hostname | IP |
| CentOS 7.6 | k8s-master | 138.138.82.14 |
| CentOS 7.6 | k8s-node1 | 138.138.82.15 |
| CentOS 7.6 | k8s-node2 | 138.138.82.16 |
注：以下是简单的安装部署操作，实际用于生产环境的话还要根据不同的后端存储进而修改yaml中的相应存储路径。
操作步骤：
1. 获取yaml文件
从github官网下载：https://github.com/kubernetes/kubernetes/tree/master/cluster/addons/fluentd-elasticsearch
解压后yanl文件位置： kubernetes/cluster/addons/fluentd-elasticsearch/
复制yaml文件到另一处进行修改操作：
[root@k8s-master ~]# mkdir efk [root@k8s-master ~]# cp kubernetes/cluster/addons/fluentd-elasticsearch/*.yaml efk/ [root@k8s-master ~]# cd efk/ [root@k8s-master efk]# ll 总用量 36 -rw-r--r-- 1 root root 382 5月 5 13:55 es-service.yaml -rw-r--r-- 1 root root 2906 5月 5 14:44 es-statefulset.yaml -rw-r--r-- 1 root root 16124 5月 5 13:55 fluentd-es-configmap.yaml -rw-r--r-- 1 root root 2702 5月 5 13:55 fluentd-es-ds.yaml -rw-r--r-- 1 root root 1052 5月 5 13:55 kibana-deployment.yaml -rw-r--r-- 1 root root 354 5月 5 13:55 kibana-service.yaml
2. 准备镜像（在所有主机上）
注：因为被墙的缘故，加上镜像文件比较大，启动会费时，可以提前单独下载好所需镜像。
①修改 es-statefulset.yaml （https://www.docker.elastic.co/#）
②修改 fluentd-es-ds.yaml
③无需修改 kibana-deployment.yaml
说明：关于kibana镜像以-oss结尾的版本好奇（default release that contains the commercially licensed code. The release that contains only open source will have the -oss appended to it.）：
https://www.elastic.co/guide/en/kibana/6.0/docker.html
附上镜像拉取文本：
~]# docker pull docker.elastic.co/elasticsearch/elasticsearch:6.6.2 ~]# docker pull registry.cn-hangzhou.aliyuncs.com/google_containers/fluentd-elasticsearch:v2.4.0 ~]# docker pull docker.elastic.co/kibana/kibana-oss:6.6.1
3. 发布yaml文件到k8s中
[root@k8s-master efk]# kubectl apply -f .
4. 查看各pod状态
[root@k8s-master efk]# kubectl get po -n kube-system -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES calico-node-b78m4 1/1 Running 0 7d1h 138.138.82.16 k8s-node2 <none> <none> calico-node-r5mlj 1/1 Running 0 7d1h 138.138.82.14 k8s-master <none> <none> calico-node-z5tdh 1/1 Running 0 7d1h 138.138.82.15 k8s-node1 <none> <none> coredns-fb8b8dccf-6mgks 1/1 Running 0 7d1h 192.168.0.7 k8s-master <none> <none> coredns-fb8b8dccf-cbtlx 1/1 Running 0 7d1h 192.168.0.6 k8s-master <none> <none> elasticsearch-logging-0 1/1 Running 0 10m 192.168.2.11 k8s-node2 <none> <none> elasticsearch-logging-1 1/1 Running 0 10m 192.168.1.17 k8s-node1 <none> <none> etcd-k8s-master 1/1 Running 0 7d1h 138.138.82.14 k8s-master <none> <none> fluentd-es-v2.4.0-fm9n2 1/1 Running 0 10m 192.168.1.16 k8s-node1 <none> <none> fluentd-es-v2.4.0-xjx6h 1/1 Running 0 10m 192.168.2.10 k8s-node2 <none> <none> kibana-logging-f4d99b69f-hc4h9 1/1 Running 0 10m 192.168.1.15 k8s-node1 <none> <none> kube-apiserver-k8s-master 1/1 Running 0 7d1h 138.138.82.14 k8s-master <none> <none> kube-controller-manager-k8s-master 1/1 Running 1 7d1h 138.138.82.14 k8s-master <none> <none> kube-proxy-c9xd2 1/1 Running 0 7d1h 138.138.82.14 k8s-master <none> <none> kube-proxy-fp2r2 1/1 Running 0 7d1h 138.138.82.15 k8s-node1 <none> <none> kube-proxy-lrsw7 1/1 Running 0 7d1h 138.138.82.16 k8s-node2 <none> <none> kube-scheduler-k8s-master 1/1 Running 1 7d1h 138.138.82.14 k8s-master <none> <none> kubernetes-dashboard-5f7b999d65-45vlk 1/1 Running 0 6d7h 192.168.1.14 k8s-node1 <none> <none> metrics-server-7579f696d8-pgcc4 1/1 Running 0 6d22h 192.168.1.12 k8s-node1 <none> <none>
顺利的话，稍等一小会就成功running状态了。
原文：https://www.cnblogs.com/ding2016/p/10815392.html
