一、准备

我这边有一个网站放在了 Google VM 上面，所以打算在购买一台服务器安装 ELK 来处理分析 nginx 的访问日志。

• 操作系统版本：CentOS 7
• ELK版本：7.1

1.1、服务器环境准备

我们这里还是采用官方的 yum 源进行安装，简单省事，首先配置官方 yum 仓库。

rpm --import https://artifacts.elastic.co/GPG-KEY-elasticsearch

cat >> /etc/yum.repos.d/elasticsearch.repo <<EOF
[elasticsearch-7.x]
name=Elasticsearch repository for 7.x packages
baseurl=https://artifacts.elastic.co/packages/7.x/yum
gpgcheck=1
gpgkey=https://artifacts.elastic.co/GPG-KEY-elasticsearch
enabled=1
autorefresh=1
type=rpm-md
EOF

关闭防火墙。

systemctl stop firewalld.service
systemctl disable firewalld.service

二、安装 ES

Elasticsearch 的软件包自带一个版本的 OpenJDK，所以我们这里不需要单独去安装 JDK 了，直接使用官方的就行。

yum install elasticsearch -y

因为默认是监听 IP 是127.0.0.1，我们需要开启配置文件/etc/elasticsearch/elasticsearch.yml的一个参数network.host，写上自己的本地IP即可。

2.1、遇到小问题

如此修改了一下配置文件之后，我发现无法启动了，日志也没有看到什么错误，我就手动执行查看问题。

[root@elk ~]# su - elasticsearch -s /bin/sh -c "/usr/share/elasticsearch/bin/elasticsearch"
su: warning: cannot change directory to /nonexistent: No such file or directory
OpenJDK 64-Bit Server VM warning: Option UseConcMarkSweepGC was deprecated in version 9.0 and will likely be removed in a future release.
[2019-05-31T03:20:42,902][INFO ][o.e.e.NodeEnvironment    ] [elk] using [1] data paths, mounts [[/ (rootfs)]], net usable_space [7.7gb], net total_space [9.9gb], types [rootfs]
[2019-05-31T03:20:42,908][INFO ][o.e.e.NodeEnvironment    ] [elk] heap size [990.7mb], compressed ordinary object pointers [true]
.................................
[2019-05-31T03:20:50,083][INFO ][o.e.x.s.a.s.FileRolesStore] [elk] parsed [0] roles from file [/etc/elasticsearch/roles.yml]
[2019-05-31T03:20:50,933][INFO ][o.e.x.m.p.l.CppLogMessageHandler] [elk] [controller/5098] [Main.cc@109] controller (64 bit): Version 7.1.1 (Build fd619a36eb77df) Copyright (c) 2019 Elasticsearch BV
[2019-05-31T03:20:51,489][DEBUG][o.e.a.ActionModule       ] [elk] Using REST wrapper from plugin org.elasticsearch.xpack.security.Security
[2019-05-31T03:20:51,881][INFO ][o.e.d.DiscoveryModule    ] [elk] using discovery type [zen] and seed hosts providers [settings]
[2019-05-31T03:20:53,019][INFO ][o.e.n.Node               ] [elk] initialized
[2019-05-31T03:20:53,020][INFO ][o.e.n.Node               ] [elk] starting ...
[2019-05-31T03:20:53,178][INFO ][o.e.t.TransportService   ] [elk] publish_address {10.154.0.3:9300}, bound_addresses {10.154.0.3:9300}
[2019-05-31T03:20:53,187][INFO ][o.e.b.BootstrapChecks    ] [elk] bound or publishing to a non-loopback address, enforcing bootstrap checks
ERROR: [2] bootstrap checks failed
[1]: max file descriptors [4096] for elasticsearch process is too low, increase to at least [65535]
[2]: the default discovery settings are unsuitable for production use; at least one of [discovery.seed_hosts, discovery.seed_providers, cluster.initial_master_nodes] must be configured
[2019-05-31T03:20:53,221][INFO ][o.e.n.Node               ] [elk] stopping ...
[2019-05-31T03:20:53,243][INFO ][o.e.n.Node               ] [elk] stopped
[2019-05-31T03:20:53,244][INFO ][o.e.n.Node               ] [elk] closing ...
[2019-05-31T03:20:53,259][INFO ][o.e.n.Node               ] [elk] closed
[2019-05-31T03:20:53,264][INFO ][o.e.x.m.p.NativeController] [elk] Native controller process has stopped - no new native processes can be started

我们可以看到下面的两个错误，一个是说最大文件打开数比较小，一个说还必须开启一个 discovery 设置，那我们分别操作一下。

#临时修改，立刻生效
ulimit -n 655350
#永久修改
echo "* soft nofile 655360" >> /etc/security/limits.conf
echo "* hard nofile 655360" >> /etc/security/limits.conf

我修改后的配置文件如下：

node.name: node-1
path.data: /var/lib/elasticsearch
path.logs: /var/log/elasticsearch
network.host: 10.154.0.3
cluster.initial_master_nodes: ["node-1", "node-2"]

然后我们就可以启动了。

systemctl daemon-reload
systemctl enable elasticsearch.service
systemctl start elasticsearch.service

访问一下页面。

{"name" : "node-1","cluster_name" : "elasticsearch","cluster_uuid" : "P0GGZU9pRzutp-zHDZvYuQ","version" : {"number" : "7.1.1","build_flavor" : "default","build_type" : "rpm","build_hash" : "7a013de","build_date" : "2019-05-23T14:04:00.380842Z","build_snapshot" : false,"lucene_version" : "8.0.0","minimum_wire_compatibility_version" : "6.8.0","minimum_index_compatibility_version" : "6.0.0-beta1"},"tagline" : "You Know, for Search"
}

三、安装 Kibana

为了节省机器资源，我们这里使用安装在一台机器上面，因为环境上面已经配置好，我们这里直接进行安装。

yum install kibana -y

编辑配置文件/etc/kibana/kibana.yml，修改如下内容：

server.port: 5601
server.host: "10.154.0.3"
elasticsearch.hosts: ["http://10.154.0.3:9200"]

启动服务。

systemctl daemon-reload
systemctl enable kibana.service
systemctl start kibana.service

访问画面正常显示即我们安装完成。

四、安装 Logstash

yum install logstash -y

创建配置文件/etc/logstash/conf.d/logstash.conf，增加如下内容：

input {beats {port => 5044}
}filter {if [fields][type] == "nginx_access" {grok {match => { "message" => "%{NGINXACCESS}" }}}if [fields][type] == "apache_access" {grok {match => { "message" => "%{COMBINEDAPACHELOG}" }}}  mutate {  remove_field =>["message"]remove_field =>["host"]remove_field =>["input"]remove_field =>["prospector"]remove_field =>["beat"]}geoip {source => "clientip"}date {match => [ "timestamp" , "dd/MMM/yyyy:HH:mm:ss Z" ]}
}output {if [fields][type] == "nginx_access" {elasticsearch {hosts => "172.18.8.200:9200"index => "%{[fields][type]}-%{+YYYY.MM.dd}"}}if [fields][type] == "apache_access" {elasticsearch {hosts => "172.18.8.200:9200"index => "%{[fields][type]}-%{+YYYY.MM.dd}"}}
}

未完待续。。。

最终效果图如下：