ELK集群部署---LogStash，Filebeat的部署

1.  环境规划:

2.  node4节点已经安装jdk：

[root@node4 ~]# java -version
java version "1.8.0_202"
Java(TM) SE Runtime Environment (build 1.8.0_202-b08)
Java HotSpot(TM) 64-Bit Server VM (build 25.202-b08, mixed mode)

3.  安装LogStash和Filebeat：

[root@node4 ~]# yum localinstall -y logstash-7.2.0.rpm

[root@node4 ~]# yum localinstall -y filebeat-7.2.0-x86_64.rpm

4.  配置Filebeat：

[root@node4 ~]# cd /etc/filebeat/

[root@node4 filebeat]# cp filebeat.yml{,.bak}

##配置filebeat，从日志文件输入，输出到kafka
[root@node4 filebeat]# vim filebeat.yml

filebeat.inputs:
  - type: log
    enabled: true
    paths:
      - /var/log/httpd/access_log
    fields:
      type: httpd-access
    multiline.pattern: ^\[
    multiline.negate: true
    multiline.match: after

  - type: log
    enabled: true
    paths:
      - /var/log/httpd/error_log
    fields:
      type: httpd-error
    multiline.pattern: ^\[
    multiline.negate: true
    multiline.match: after

  - type: log
    enabled: true
    paths:
      - /var/log/mariadb/mariadb.log
    fields:
      type: mariadb
    multiline.pattern: ^\[
    multiline.negate: true
    multiline.match: after

output.kafka:
  hosts: ["192.168.56.111:9092","192.168.56.112:9092","192.168.56.113:9092"]  ##kafka集群节点
  topic: "%{[fields][type]}"  ##kafka的标题

multiline参数解析：1.  " multiline.type "：默认值是pattern。

                                2. " multiline.pattern "：匹配日志的正则，一般情况下会配置为从行首开始匹配。

                                3. " multiline.match，multiline.negate "：multiline.match追加的位置， multiline.negate是否为否定模式。

multiline.match为after，multiline.negate为true的意思就是从匹配的位置开始，只要是不满足匹配条件的，全部追加到匹配的那一行，直到再次遇到匹配条件的位置。

5.  配置logstash：

[root@node4 ~]# cd /etc/logstash/conf.d

##配置logstash：输入为kafka，输出为elasticsearch。
[root@node4 conf.d]# vim all.conf

input{
  kafka {
    bootstrap_servers => "192.168.56.111:9092,192.168.56.112:9092,192.168.56.113:9092"
    codec => json
    topics => "httpd-access"  ##匹配kafka中的主题
    consumer_threads => 1
    decorate_events => true
    type => "httpd-access"  ##用于输出时条件判断
  }

  kafka {
    bootstrap_servers => "192.168.56.111:9092,192.168.56.112:9092,192.168.56.113:9092"
    codec => json
    topics => "httpd-error"  ##匹配kafka中的主题
    consumer_threads => 1
    decorate_events => true
    type => "httpd-error"  ##用于输出时条件判断
  }

  kafka {
    bootstrap_servers => "192.168.56.111:9092,192.168.56.112:9092,192.168.56.113:9092"
    codec => json
    topics => "mariadb"  ##匹配kafka中的主题
    consumer_threads => 1
    decorate_events => true
    type => "mariadb"  ##用于输出时条件判断
  }

}

##输出时如果满足type的判断条件，就按照指定索引输出到elasticsearch上。
output {
  if [type] == "httpd-access" {
    elasticsearch {
      hosts => ["192.168.56.111:9200","192.168.56.112:9200","192.168.56.113:9200"]
      index => "httpd-accesslog-%{+yyyy.MM.dd}"
    }
  }

  if [type] == "httpd-error" {
    elasticsearch {
      hosts => ["192.168.56.111:9200","192.168.56.112:9200","192.168.56.113:9200"]
      index => "httpd-errorlog-%{+yyyy.MM.dd}"
    }
  }

  if [type] == "mariadb" {
    elasticsearch {
      hosts => ["192.168.56.111:9200","192.168.56.112:9200","192.168.56.113:9200"]
      index => "mariadblog-%{+yyyy.MM.dd}"
    }
  }
}

6.  测试Filebeat和LogStash是否可以成功采集到日志。

[root@node4 ~]# cd /etc/filebeat

[root@node4 filebeat]# filebeat -e -c filebeat.yml

另起一个终端启动logstash：
[root@node4 ~]# cd /etc/logstash/conf.d

[root@node4 conf.d]# logstash -f all.conf

访问apache http和登录mariadb，查看elasticsearch上是否采集到指定索引的日志。

如果想以服务方式启动filebeat和logstash，filebeat配置文件名就要是filebeat.yml;logstash配置文件必须要在/etc/filebeat/conf.d目录下，并且后缀为".conf"。 

[root@node4 ~]# systemctl start filebeat.service

[root@node4 ~]# systemctl start logstash.service