ELK5.3+Kafka集群配置

【一】资源准备

# 3台4C*8G, 安装Zookeeper、Kafka、Logstash--Broker(input: filebeat; output: Kafka)

10.101.2.23 10.101.2.24 10.101.2.25

# 2台4C*8G, 安装Logstash--Indexer(input: Kafaka; output: Elasticsearch)

10.101.2.26 10.101.2.27

# 3台8C*16G, 安装Elasticsearch

10.101.2.28 10.101.2.29 10.101.2.30

# 2台2C*4G, 安装Kibana

10.101.2.31 10.101.2.32

# 安装包下载

elasticsearch-5.3.1.tar.gz

filebeat-5.3.1-linux-x86_64.tar.gz

jdk-8u131-linux-x64.tar.gz

kafka_2.12-0.10.2.0.tgz

kibana-5.3.1-linux-x86_64.tar.gz

logstash-5.3.1.tar.gz

node-v7.9.0-linux-x64.tar.gz

zookeeper-3.4.10.tar.gz

nginx-1.12.0.tar.gz

统一上传至服务器 /usr/local/src 目录下

【二】通用配置

# 配置hosts

vi /etc/hosts

10.101.2.23 vmserver2x23

10.101.2.24 vmserver2x24

10.101.2.25 vmserver2x25

10.101.2.26 vmserver2x26

10.101.2.27 vmserver2x27

10.101.2.28 vmserver2x28

10.101.2.29 vmserver2x29

10.101.2.30 vmserver2x30

10.101.2.31 vmserver2x31

10.101.2.32 vmserver2x32

# 配置ssh访问限制，假如有必要的话

vi /etc/hosts.allow

【三】安装Elasticsearch集群

# 系统环境

vi /etc/sysctl.conf

vm.max_map_count=262144

fs.file-max=65536

执行 sysctl -p 使配置生效

vi /etc/security/limits.conf #打开文件数

* soft nofile 65536

* hard nofile 131072

* soft nproc 2048

* hard nproc 4096

* - memlock unlimited

vi /etc/security/limits.d/90-nproc.conf

* soft nproc 2048

# 配置Java环境变量

cd /usr/local/src/

tar -xvf jdk-8u131-linux-x64.tar.gz

mv jdk1.8.0_131 /usr/share/

vi /etc/profile #在末尾添加下面3行后, 保存退出

export JAVA_HOME=/usr/share/jdk1.8.0_131

export PATH=$JAVA_HOME/bin:$PATH

export CLASSPATH=.:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar

source /etc/profile 使得配置立即生效

# 解压elasticsearch

cd /usr/local/src

tar -xvf elasticsearch-5.3.1.tar.gz

mv elasticsearch-5.3.1 /usr/local

vi /usr/local/elasticsearch-5.3.1/bin/elasticsearch #设置ES_JAVA_OPTS参数

ES_JAVA_OPTS="-Xms8g -Xmx8g" # 去掉该行前的注释后, 一定要删除后面的这一串 ./bin/elasticsearch

# 新增elastic组及用户, 因为elasticsearch不允许root用户启动

groupadd elastic

useradd elastic -g elastic

passwd elastic # 设定用户密码

chown -R elastic:elastic /usr/local/elasticsearch-5.3.1/

# 配置elasticsearch.ywl, 主要参数如下

cluster.name: bsd-elk

node.name: elk-2-30 # 每个节点不同

node.master: true

node.data: true

bootstrap.memory_lock: true

bootstrap.system_call_filter: false # centos7以下版本需要将这个参数设置为false

network.host: 0.0.0.0

http.port: 9200

discovery.zen.ping.unicast.hosts: ["10.101.2.28:9300", "10.101.2.29:9300", "10.101.2.30:9300"]

discovery.zen.minimum_master_nodes: 2

discovery.zen.ping_timeout: 60s # 网上大部分文章这个参数都写成了discovery.zen.ping.timeout

http.cors.enabled: true

http.cors.allow-origin: "*"

# 下载node-v7.9.0-linux-x64.tar.gz, 解压后mv到/usr/local/nodejs-7.9.0

chown -R elastic:elastic nodejs-7.9.0/

cd /usr/local/nodejs-7.9.0

ln -s /usr/local/nodejs-7.9.0/bin/node /usr/local/bin

ln -s /usr/local/nodejs-7.9.0/bin/npm /usr/local/bin

# 安装head插件, 5.x以上的elasticsearch暂没有找到离线安装的方法, 所以需要服务器开通外网访问权限

# 5.x版本是里程碑式的更新, 网上大部分的文章都是以前版本的插件安装

cd /usr/local/elasticsearch-5.3.1

git clone https://github.com/mobz/elasticsearch-head.git

如果没有git工具先安装, yum install git

cd elasticsearch-head

npm install -g grunt --registry=https://registry.npm.taobao.org # 安装grunt

npm install # 安装head

npm install grunt --save # 如果 node_modules/grunt/bin/目录下没有 grunt文件, 就执行以下该命令

vi Gruntfile.js 修改connect, 在options里添加本机IP hostname: '10.101.2.30',

cd /usr/local/elasticsearch-5.3.1

bin/elasticsearch -d # 启动elasticsearch

cd elasticsearch-head

node_modules/grunt/bin/grunt server & # 启动head插件

访问 http://10.101.2.30:9100

# 安装bigdesk插件

cd /usr/local/elasticsearch-5.3.1

git clone https://github.com/hlstudio/bigdesk

cd bigdesk/_site

python -m SimpleHTTPServer & # 启动bigdesk插件

访问 http://10.101.2.30:8000

另外两台机器(10.101.2.28 10.101.2.29)按此步骤同样配置, master和data节点的选择看实际情况, 我的3台全是混合节点

所有elasticsearch启动后, 访问head如果能看到3个节点的集群信息, 就可以了

【四】安装ZooKeeper集群

# zookeeper依赖java, java环境配置参照上面

# 解压zookeeper-3.4.10.tar.gz

cd /usr/local/src

tar -xvf zookeeper-3.4.10.tar.gz

mv zookeeper-3.4.10 /usr/local

mkdir /usr/local/zookeeper-3.4.10/data # 每个节点上创建一个数据存放目录

# 创建myid文件

echo 23 >/usr/local/zookeeper-3.4.10/data/myid # 10.101.2.23、24、25三台机器上myid的值分别放 23 24 25

# 配置zoo.cfg

cd /usr/local/zookeeper-3.4.10/conf/

cp zoo_sample.cfg zoo.cfg

vi zoo.cfg # 主要参数如下

tickTime=2000

initLimit=10

syncLimit=5

dataDir=/usr/local/zookeeper-3.4.10/data

clientPort=2181

server.23=10.101.2.23:2888:3888

server.24=10.101.2.24:2888:3888

server.25=10.101.2.25:2888:3888

# 复制配置文件到其他节点

scp zoo.cfg root@ip:/usr/local/zookeeper-3.4.10/conf/

# 启动zookeeper集群

cd /usr/local/zookeeper-3.4.10/

bin/zkServer.sh start

bin/zkServer.sh status # 主节点会返回 Mode: leader, 从节点返回 Mode: follower

至此zookeeper集群配置完毕

【五】配置kafka集群

# 解压kafka_2.12-0.10.2.0.tgz, 创建数据目录

cd /usr/local

tar -xvf src/kafka_2.12-0.10.2.0.tgz

mkdir /usr/local/kafka_2.12-0.10.2.0/data

# 配置server.propertites

cd /usr/local/kafka_2.12-0.10.2.0/config

vi server.properties # 主要参数如下

broker.id=23 # 10.101.2.23、24、25三台机器上id的值分别放 23 24 25

delete.topic.enable=true

num.network.threads=3

num.io.threads=8

socket.send.buffer.bytes=102400

socket.receive.buffer.bytes=102400

socket.request.max.bytes=104857600

log.dirs=/usr/local/kafka_2.12-0.10.2.0/data

num.partitions=6

num.recovery.threads.per.data.dir=1

#log.flush.interval.messages=10000

#log.flush.interval.ms=1000

log.retention.hours=72

#log.retention.bytes=1073741824

log.segment.bytes=1073741824

log.retention.check.interval.ms=300000

zookeeper.connect=10.101.2.23:2181,10.101.2.24:2181,10.101.2.25:2181

zookeeper.connection.timeout.ms=6000

# 复制配置文件到其他节点, 不要忘记修改 broker.id

scp server.properties root@ip:/usr/local/kafka_2.12-0.10.2.0/config/

# 启动kafka集群

cd /usr/local/kafka_2.12-0.10.2.0/

bin/kafka-server-start.sh config/server.properties > /dev/null &

# 友情赠送几条常用命令

bin/kafka-topics.sh --create --zookeeper localhost:2181 --replication-factor 1 --partitions 1 --topic test # 创建topic

bin/kafka-topics.sh --list --zookeeper localhost:2181 # 查看已经创建的topic列表

bin/kafka-topics.sh --describe --zookeeper localhost:2181 --topic test # 查看topic的详细信息

bin/kafka-console-producer.sh --broker-list localhost:9092 --topic test # 发送消息, 回车后模拟输入一下消息

bin/kafka-console-consumer.sh --bootstrap-server localhost:9092 --topic test # 消费消息, 可以换到其他kafka节点, 同步接收生产节点发送的消息

bin/kafka-topics.sh --zookeeper localhost:2181 --alter --topic test --partitions 6 # 给topic增加分区

bin/kafka-topics.sh --delete --zookeeper localhost:2181 --topic test1 # 删除已经创建的topic, 前提是开了delete.topic.enable=true参数

如果还不能删除, 可以到zookeeper中去干掉它

cd /usr/local/zookeeper-3.4.10/

bin/zkCli.sh

ls /brokers/topics # 查看topic

rmr /brokers/topics/test1 # 删除topic

至此kafka集群配置完毕

【六】logstash--broker集群配置

# java环境配置略

# 解压logstash-5.3.1.tar.gz

cd /usr/local

tar -xvf src/logstash-5.3.1.tar.gz

# 添加配置文件 beat_to_kafka.conf

cd logstash-5.3.1

vi config/beat_to_kafka.conf # 输入下面内容后保存

input {

beats {

port => 5044

}

}

filter {

}

# topic_id改成按beat中配置的document_type来输出到不同的topic中, 供kibana分组过滤用

output {

kafka {

bootstrap_servers => "10.101.2.23:9092,10.101.2.24:9092,10.101.2.25:9092"

# topic_id => "bsd-log"

topic_id => '%{[type]}'

}

}

# 启动logstash

bin/logstash -f config/beat_to_kafka.conf > /dev/null &

至此logstash-broker集群配置完成

【七】在实际应用服务器上安装filebeat

# 解压filebeat-5.3.1-linux-x86_64.tar.gz

cd /usr/local/

tar -xvf src/filebeat-5.3.1-linux-x86_64.tar.gz

mv filebeat-5.3.1-linux-x86_64 filebeat-5.3.1

# 配置filebeat.yml文件, 先找一台drds试水

cd filebeat-5.3.1

vi filebeat.yml # 主要参数如下

#=========================== 文件正文开始 =============================

filebeat.prospectors:

-

input_type: log

paths:

- /home/admin/drds-server/3306/logs/rms/slow.log

- /home/admin/drds-server/3306/logs/engineering/slow.log

- /home/admin/drds-server/3306/logs/sc_file/slow.log

- /home/admin/drds-server/3306/logs/sc_user/slow.log

- /home/admin/drds-server/3306/logs/sc_order/slow.log

- /home/admin/drds-server/3306/logs/sc_inventory/slow.log

- /home/admin/drds-server/3306/logs/sc_marketing/slow.log

- /home/admin/drds-server/3306/logs/sc_message/slow.log

- /home/admin/drds-server/3306/logs/sc_channel/slow.log

#exclude_lines: ["^DBG"]

#include_lines: ['Exception','ERR_CODE']

#exclude_files: [".gz$"]

document_type: drds-slow

# 设定合并正则

multiline.pattern: ^[0-9]{4}-[0-9]{2}-[0-9]{2}\s[0-9]{2}:[0-9]{2}:[0-9]{2}.[0-9]{3}

multiline.negate: true

multiline.match: after

# 单台机器上配置不同document_type

-

input_type: log

paths:

- /home/admin/drds-server/3306/logs/test/sql.log

document_type: drds-sql

multiline.pattern: ^[0-9]{4}-[0-9]{2}-[0-9]{2}\s[0-9]{2}:[0-9]{2}:[0-9]{2}.[0-9]{3}

multiline.negate: true

multiline.match: after

#----------------------------- Logstash output --------------------------------

output.logstash:

# The Logstash hosts

hosts: ["10.101.2.23:5044","10.101.2.24:5044","10.101.2.25:5044"]

#=========================== 文件正文结束 =============================

# 启动filebeat

./filebeat -c filebeat.yml > /dev/null &

【八】配置logstash--indexer集群

# java环境配置略

# 解压logstash-5.3.1.tar.gz

cd /usr/local

tar -xvf src/logstash-5.3.1.tar.gz

# 添加配置文件 kafka_to_es.conf

cd logstash-5.3.1

vi config/kafka_to_es.conf # 输入下面内容后保存

# input里关于server以及topic的配置, 5.x和以前版本不同

input {

kafka {

bootstrap_servers => "10.101.2.23:9092,10.101.2.24:9092,10.101.2.25:9092"

group_id => "logstash"

topics => ["drds-sql","drds-slow","sc_user","sc_channel","sc_order","sc_inventory","sc_message","sc_file","sc_marketing","rms",'scm','engineering']

consumer_threads => 50

decorate_events => true

}

}

filter {

}

output {

elasticsearch {

hosts => ["10.101.2.28:9200","10.101.2.29:9200","10.101.2.30:9200"]

index => "logstash-%{+YYYY.MM.dd.hh}"

manage_template => true

template_overwrite => true

template_name => "drdsLogstash"

flush_size => 50000

idle_flush_time => 10

}

}

# 启动logstash

bin/logstash -f config/kafka_to_es.conf > /dev/null &

至此logstash-indexer集群配置完成, 不出意外, elasticsearch-head里应该已经有数据写入了

【九】配置kibana集群

# 解压kibana-5.3.1-linux-x86_64.tar.gz

cd /usr/local

tar -xvf src/kibana-5.3.1-linux-x86_64.tar.gz

mv kibana-5.3.1-linux-x86_64/ kibana-5.3.1

# 配置kibana.yml文件

cd kibana-5.3.1

vi config/kibana.yml #主要参数如下

server.port: 5601

server.host: "0.0.0.0"

elasticsearch.url: "http://10.101.2.28:9200" # 指到es集群master节点

# 启动kibana

bin/kibana > /dev/null &

# kibana对浏览器兼容性不好, 低版本的chrome和ie都不能访问, 显示正在加载

访问 http://ip:5601

# 另一台kibana节点按相同配置(可以把es.url指到另一个节点), kibana查询支持布尔运算符, 通配符等, 关键字要大些(如AND OR), 具体自行百度

【十】配置nginx代理

# 安装nginx需要的一些依赖包

yum -y install pcre-devel

yum -y install gd-devel

# 解压nginx-1.12.0.tar.gz

cd /usr/local/

tar -xvf src/nginx-1.12.0.tar.gz

# 安装nginx

cd nginx-1.12.0

./configure --prefix=/usr/local/nginx-1.12.0/ --conf-path=/usr/local/nginx-1.12.0/nginx.conf

make

make install

# 配置 nginx.conf 文件, 我们这边只做负载均衡, 随便设设就好

vi /usr/local/nginx-1.12.0/nginx.conf

worker_processes 1;

error_log logs/error.log info;

#pid logs/nginx.pid;

events {

worker_connections 1024;

}

http {

include mime.types;

default_type application/octet-stream;

log_format main '$remote_addr - $remote_user [$time_local] "$request" '

'$status $body_bytes_sent "$http_referer" '

'"$http_user_agent" "$http_x_forwarded_for"';

access_log logs/access.log main;

sendfile on;

keepalive_timeout 65;

upstream kibana {

server 10.101.2.31:5601;

server 10.101.2.32:5601;

}

server {

listen 15601;

server_name 10.101.2.31;

#charset koi8-r;

#access_log logs/host.access.log main;

location / {

root html;

index index.html index.htm;

proxy_pass http://kibana;

}

error_page 500 502 503 504 /50x.html;

location = /50x.html {

root html;

}

}

# 启动nginx

sbin/nginx

# 之后在浏览器上访问 http://nginx_ip:15601 即可

至此集群所有组件配置完成