企业级ELKF物理机集群搭建
企业级架构ELKF集群
服务器
A:10.10.1.3
B:10.10.1.4
C:10.10.1.5
客户端
D:10.10.1.6
下面将用ABC代表服务器了
ABC修改计算机名称
A服务器:hostnamectl set-hostname node1B服务器:hostnamectl set-hostname node2C服务器:hostnamectl set-hostname node3
ADCD更新源并安装java
curl -o /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repocurl -o /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repoyum install lrzsz vim java-1.8.0-openjdk java-1.8.0-openjdk-devel-y java –version -y
ABC都切换到
cd /usr/local/src/
吧文件下面文件都拷贝到上面的这个目录
apache-zookeeper-3.6.0-bin.tar.gzelasticsearch-7.13.1-x86_64.rpmfilebeat-7.13.1-x86_64.rpmkafka_2.12-2.5.0.tgzkibana-7.13.1-x86_64.rpmlogstash-7.13.1-x86_64.rpm
ABC都要进行更新源,和安装java
curl -o /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repocurl -o /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repoyum install lrzsz vim java-1.8.0-openjdk java-1.8.0-openjdk-devel-y java –version -y
ABC安装ES集群服务,进行连接
cd /usr/local/src/yum localinstall elasticsearch-7.13.1-x86_64.rpm -y
JVM的内存限制更改,根据实际环境调整,更具服务器内存修改,如果服务器8G,这里可以改为6G,给服务器预留1-2G即可
vi /etc/elasticsearch/jvm.options -Xms4g 改为 -Xmx2g
ES集群实战注意
集群交互是使用证书交互
搭建集群前需要先创建证书文件
ES集群交互CA证书创建 A服务器上
/usr/share/elasticsearch/bin/elasticsearch-certutil ca #一直回车
查看证书位置
ll -rht /usr/share/elasticsearch/elastic-stack-ca.p12
创建证书
/usr/share/elasticsearch/bin/elasticsearch-certutil cert --ca /usr/share/elasticsearch/elastic-stack-ca.p12 #一直回车
查看证书
ll -rht /usr/share/elasticsearch/elastic-certificates.p12cp /usr/share/elasticsearch/elastic-certificates.p12 /etc/elasticsearch/elastic-certificates.p12
证书要求权限修改,不然集群搭建失败
chmod 600 /etc/elasticsearch/elastic-certificates.p12chown elasticsearch:elasticsearch /etc/elasticsearch/elastic-certificates.p12
吧加密ca证书拷贝到BC服务器
scp /etc/elasticsearch/elastic-certificates.p12 10.10.1.4:/etc/elasticsearch/scp /etc/elasticsearch/elastic-certificates.p12 10.10.1.5:/etc/elasticsearch/
ABC服务器对比MD5值是否一致
md5sum /etc/elasticsearch/elastic-certificates.p12
B服务器和C服务器都要执行
chmod 600 /etc/elasticsearch/elastic-certificates.p12chown elasticsearch:elasticsearch /etc/elasticsearch/elastic-certificates.p12
ADC服务器都要进行修改
vi /etc/elasticsearch/elasticsearch.ymlcluster.name: maixiaolunnode.name: node1 #服务器A就改为nede1,服务器B改为node2,服务器C改为node3node.master: truenode.data: truepath.data: /var/lib/elasticsearchpath.logs: /var/log/elasticsearchnetwork.host: 0.0.0.0http.port: 9200discovery.seed_hosts: ["10.10.1.3", "10.10.1.4", "10.10.1.5"]cluster.initial_master_nodes: ["10.10.1.3", "10.10.1.4", "10.10.1.5"]xpack.security.enabled: truexpack.monitoring.enabled: truexpack.security.transport.ssl.enabled: truexpack.security.transport.ssl.verification_mode: certificatexpack.security.transport.ssl.keystore.path: /etc/elasticsearch/elastic-certificates.p12xpack.security.transport.ssl.truststore.path: /etc/elasticsearch/elastic-certificates.p12
ABC防火墙要放通端口
firewall-cmd --zone=public --add-port=9300/tcp --permanentfirewall-cmd --zone=public --add-port=9300/udp --permanentfirewall-cmd --zone=public --add-port=9200/tcp --permanentfirewall-cmd --zone=public --add-port=9200/udp --permanentfirewall-cmd --zone=public --add-port=5601/tcp --permanentfirewall-cmd --zone=public --add-port=5601/udp --permanentfirewall-cmd --zone=public --add-port=80/tcp --permanentfirewall-cmd --zone=public --add-port=3000/tcp --permanent
重新载入
firewall-cmd --reload
ABC服务器顺序启动ES服务
systemctl enable elasticsearchsystemctl restart elasticsearch
继续看第一台服务器日志
出现valid,表示成功了
接下来是查看3台集群中间的互动
yum -y install net-tools 安装 netstatnetstat -anp |grep 10.0.0.20:9300 查看集群之间的交互
检查9300端口连接情况
确认集群中所有es的日志正常再设置密码maixiaolun..123
在主服务器(服务器A,想把那个当主服务器就在那个服务器上运行)上运行
ES设置密码/usr/share/elasticsearch/bin/elasticsearch-setup-passwords interactive
ES设置随机密码/usr/share/elasticsearch/bin/elasticsearch-setup-passwords auto
验证集群是否成功,标记为*的为master节点。网页访问或者curl访问
http://10.10.1.3:9200 账号:elastic 密码:刚刚设置的
http://10.10.1.3:9200/_cat/nodes?v //查看节点信息 带*的就是主的
http://xxx:9200/_cat/indices?v
命令方式查看 输入账号密码
curl -u elastic:maixiaolun..123 http://10.10.1.3:9200/_cat/nodes?v
B服务器安装Kibana (因为是前端,蹦了就蹦了,不要紧,所有不需要每个地方都搭建来)
B服务器安装
cd /usr/local/src/yum localinstall kibana-7.13.1-x86_64.rpm -y
Kibana配置连接ES集群
vi /etc/kibana/kibana.ymlserver.port: 5601server.host: "0.0.0.0"elasticsearch.hosts: ["http://10.10.1.3:9200", "http://10.10.1.4:9200", "http://10.10.1.5:9200"]elasticsearch.username: "elastic"elasticsearch.password: "maixiaolun..123"logging.dest: /tmp/kibana.log
B服务器Kibana的启动和访问
systemctl enable kibanasystemctl restart kibana
Kibana监控开启
netstat -tulnp
http://10.10.1.3:5601
账号:elastic
密码:maixiaolun..123
BC服务器安装logstash日志分析 (就是蹦了一台也不要紧,此做法为高可用做法,安装一台的话也是可以的)
cd /usr/local/src/yum localinstall logstash-7.13.1-x86_64.rpm -y
内存配置jvm.options 1g,根据自己实际情况改
vi /etc/logstash/jvm.options
配置logstash配置文件
vi /etc/logstash/conf.d/logstash.conf
#监听5044传来的日志
input { beats {
host => '0.0.0.0'
port => 5044
}
}
#这里是正则表达式,分析日志
filter {
grok {
match => {
"message" => '%{IP:remote_addr} - (%{WORD:remote_user}|-) [%
{HTTPDATE:time_local}] "%{WORD:method} %{NOTSPACE:request} HTTP/%
{NUMBER}" %{NUMBER:status} %{NUMBER:body_bytes_sent} %{QS} %
{QS:http_user_agent}'
}
remove_field => ["message"]
}
date {
match => ["time_local", "dd/MMM/yyyy:HH:mm:ss Z"]
target => "@timestamp"
}
}
#发送给所有ES服务器
output {
elasticsearch {
hosts => ["http://10.10.1.3:9200", "http://10.10.1.4:9200", "http://10.10.1.5:9200"]
user => "elastic"
password => "maixiaolun..123"
index => "mxlnginx-%{+YYYY.MM.dd}"
}
}
启动Logstash,配置重载:kill -1 pid
systemctl restart logstashsystemctl enable logstash
等待启动,查看日志
/var/log/logstash 日志生成的地方
tail -f logstash-plain.log
ZK集群部署
所有服务器都要进行安装
cd /usr/local/src/tar -zxvf apache-zookeeper-3.6.0-bin.tar.gzmv apache-zookeeper-3.6.0-bin /usr/local/zookeepermkdir -pv /usr/local/zookeeper/datacd /usr/local/zookeeper/confcp zoo_sample.cfg zoo.cfg
zk集群至少需要三台机器
集群配置zoo.cfg
vi zoo.cfgtickTime=2000initLimit=10syncLimit=5dataDir=/usr/local/zookeeper/dataclientPort=2181autopurge.snapRetainCount=3autopurge.purgeInterval=1server.1=10.10.1.3:2888:3888server.2=10.10.1.4:2888:3888server.3=10.10.1.5:2888:3888
更改zk集群的id
vi /usr/local/zookeeper/data/myid
第一台就填写1
第二台就填写2
第三台就填写3
分别为1 2 3
systemctl管理
vi /usr/lib/systemd/system/zookeeper.service[Unit]Description=zookeeperAfter=network.target[Service]Type=forkingExecStart=/usr/local/zookeeper/bin/zkServer.sh startUser=root[Install]WantedBy=multi-user.target
启动zk
systemctl enable zookeepersystemctl restart zookeeper
启动zk集群查看状态
cd /usr/local/zookeeper/bin/./zkServer.sh start./zkServer.sh status
报错的话,检查防火墙。
正确显示为下
Kafka集群部署
下载地址:http://kafka.apache.org/downloads
安装
cd /usr/local/src/tar -zxvf kafka_2.12-2.5.0.tgz -C /usr/local/
Jvm内存修改/usr/local/kafka_2.12-2.5.0/bin/kafka-server-start.sh,根据实际情况修
改,每台服务器都要改
修改Kafka配置server.properties
cd /usr/local/kafka_2.12-2.5.0/config/vi server.properties
broker.id=0listeners=PLAINTEXT://xxx:9092 log.retention.hours=1 #根据实际情况修改 日志保留时间zookeeper.connect=xxx:2181,xxx:2181,xxx:2181zookeeper.connection.timeout.ms=18000 改为 zookeeper.connection.timeout.ms=60000拷贝到其他服务器上scp server.properties 10.0.0.21:/usr/local/kafka_2.12-2.5.0/config/然后在新的服务器上vi /usr/local/kafka_2.12-2.5.0/config/server.properties改listeners=PLAINTEXT://xxx:9092 对应当前服务器的ipbroker.id=0 改为broker.id=1
如果启动的时候报如下错误,是因为链接超时,吧60000改为 1200000也就是2分钟
Kafka使用systemctl管理
vi /usr/lib/systemd/system/kafka.service[Unit]Description=kafkaAfter=network.target[Service]Type=simpleExecStart=/usr/local/kafka_2.12-2.5.0/bin/kafka-server-start.sh /usr/local/kafka_2.12-2.5.0/config/server.propertiesUser=root[Install]WantedBy=multi-user.target
Kafka启动
systemctl enable kafka systemctl restart kafka
查看kafka日志
cd /usr/local/kafka_2.12-2.5.0/logstail -f server.log
创建topic,创建成功说明kafka集群搭建成功
--replication-factor 3 代表3台服务器
cd /usr/local/kafka_2.12-2.5.0/bin/./kafka-topics.sh --create --zookeeper 10.10.1.4:2181 --replication-factor 3 --partitions 1 --to
logstash获取kafka数据
vi /etc/logstash/conf.d/logstash.conf
#获取kafka的日志
input {
kafka {
bootstrap_servers => "10.10.1.3:9092,10.10.1.4:9092,10.10.1.5:9092"
topics => ["mxlkafka"]
group_id => "mxlgroup"
codec => "json"
}
}
filter {
mutate {
remove_field => ["agent","ecs","log","input","[host][os]"]
}
}
#正则表达式分析日志
filter {
grok {
match => {
"message" => '%{IP:remote_addr} - (%{WORD:remote_user}|-) [%
{HTTPDATE:time_local}] "%{WORD:method} %{NOTSPACE:request} HTTP/%
{NUMBER}" %{NUMBER:status} %{NUMBER:body_bytes_sent} %{QS} %
{QS:http_user_agent}'
}
remove_field => ["message"]
}
date {
match => ["time_local", "dd/MMM/yyyy:HH:mm:ss Z"]
target => "@timestamp"
}
}
#吧分析过后的日志发送给ES
output {
elasticsearch {
hosts => ["http://10.10.1.3:9200", "http://10.10.1.4:9200", "http://10.10.1.5:9200"]
user => "elastic"
password => "maixiaolun..123"
index => "mxlnginx-%{+YYYY.MM.dd}" #这里代表日志的名字,不通服务器可以写不通的名字
}
}
重载
ps auxfww | grep logstashkill -1 804
看日志
cd /var/log/logstash/
tail -f logstash-plain.log
测试环境是否正常
效果同ELFK,但架构做了优化
Kafka查看队列信息,验证新架构是否生效
cd /usr/local/kafka_2.12-2.5.0/bin/./kafka-consumer-groups.sh --bootstrap-server 10.10.1.4:9092 --list./kafka-consumer-groups.sh --bootstrap-server 10.10.1.4:9092 --group mxlgroup -describe
LOG-END-OFFSET不断增大,LAG不堆积说明架构生效
LOG-END-OFFSET 一直增加就代表kafka在消费
LAG 一直在堆积
Logstash扩展
配置保持一致,启动即可
客户端D进行安装nginx进行测试数据是否传入服务器集群
安装nginx做测试,如果有其他服务可以直接测试
yum install nginx -ysystemctl start nginx
安装filebeat,用于发送日志到服务器
cd /usr/local/src/yum localinstall filebeat-7.13.1-x86_64.rpm -y
配置filebeat进行监控nginx
vi /etc/filebeat/filebeat.yml
filebeat.inputs:
- type: log
tail_files: true
backoff: "1s"
paths:
- /var/log/nginx/access.log
processors:
- drop_fields:
fields: ["agent","ecs","log","input"]
output:
kafka:
hosts: ["10.10.1.3:9092", "10.10.1.4:9092", "10.10.1.5:9092"]
topic: mxlkafka #这里代表日志的名字,不通服务器可以写不通的名字
启动filebeat
systemctl start filebeatsystemctl enable filebeat
登入B服务器网页
http://10.10.1.4:5601
账号:elastic
密码:maixiaolun..123
打开Stack Management
添加日志数据 Index Patterns
添加
输入刚刚设置的名字 mxlnginx* 点击下一步
选择@timestamp 下一步
这时候就能看到添加的内容了
访问客户端地址nginx页面 10.10.1.6 多刷新几次
就可以去看数据了
看到日志已经上来了
读者可在公众号后台回复关键字 ELKF集群7搭建 获取二进制安装包