Elasticsearch cluster deploy
准备好deploy APP二进制包文件：
1 | curl -L -O https//artifacts.elastic.co/downloads/elasticsearch/elasticsearch-6.3.0.tar.gz |
创建普通用户：
1 | id deploy &>/dev/null; if [[ $? != "0" ]]; then useradd deploy; fi |
切换用户：
1 | su - deploy |
创建 APP目录和数据目录：
1 | cd ~ |
更改deploy目录和数据目录权限：
1 | chown -R deploy:deploy /home/deploy/{elasticsearch,logstash,kibana} #如果是放到单独分区的目录里注意赋权： chown -R deploy:deploy /deploy |
安装：
1 | tar xvf elasticsearch-6.3.0.tar.gz -C /home/deploy/elasticsearch |
修改配置文件：
1 | cp /home/deploy/elasticsearch/elasticsearch-6.3.0/config/elasticsearch.yml{,.ori} |
1 | cluster.name: es-cluster #自定义，集群内统一 |
后台启动elasticsearch集群
1 | su - deploy -c '/home/deploy/elasticsearch/elasticsearch-6.3.0/bin/elasticsearch -d' |
使用 ansible 将配置好的配置文件分发到各节点主机上
1 | ansible elasticsearch_servers -i elasticsearch_hosts -m file -a "path=/home/deploy/elasticsearch state=directory" ansible elasticsearch_servers -i elasticsearch_hosts -m copy -a "src=/home/deploy/elasticsearch-6.3.0.tar.gz dest=/home/deploy/ owner=root group=root mode=0644" |
使用 unarchlive 模块解压 zip 文件前需要先在目标主机上安装 unzip 包
1 | ansible elasticsearch_servers -i elasticsearch_hosts -m yum -a "name=unzip state=present" |
安装 elasticsearch-analysis-ik 插件
1 | ansible elasticsearch_servers -i elasticsearch_hosts -m unarchive -a "src=elasticsearch-analysis-ik-6.3.0.zip dest=/home/deploy/elasticsearch/elasticsearch-6.3.0/plugins/ik" |
最后再赋权一遍，以防万一：
1 | ansible elasticsearch_servers -i elasticsearch_hosts -m command -a "chown -R deploy:deploy /home/deploy/deploy" |
elasticsearch服务启停操作：
1 | ansible elasticsearch_servers -i elasticsearch_hosts -m command -a "su - deploy -c /home/deploy/elasticsearch/elasticsearch-6.3.0/bin/elasticsearch -d" |
查看集群健康状态：
curl http:/10.0.0.11:9200/_cluster/health
如果返回status=green表示正常，返回结果类似如下：
1 | {"cluster_name":"es-cluster","status":"green","timed_out":false,"number_of_nodes":2,"number_of_data_nodes":2,"active_primary_shards":0,"active_shards":0,"relocating_shards":0,"initializing_shards":0,"unassigned_shards":0,"delayed_unassigned_shards":0,"number_of_pending_tasks":0,"number_of_in_flight_fetch":0,"task_max_waiting_in_queue_millis":0,"active_shards_percent_as_number":100.0} |
查看集群状态：
http://10.0.0.11:9200/_cluster/state
http://10.0.0.11:9200/_cluster/state?pretty
FAQ：
Q:
报错：org.elasticsearch.bootstrap.StartupException: java.lang.RuntimeException: can not run elasticsearch as root
A:
不能以root账号执行，要以普通用户执行：su - deploy -c '/home/deploy/elasticsearch/elasticsearch-6.3.0/bin/elasticsearch -d'
Q：elasticsearch java.nio.file.AccessDeniedException
A：
出现这个异常的原因是当前用户没权限，解决办法，为普通用户deploy授权：
chown -R deploy:deploy /usr/local/elasticsearch
=======================================================
Q:
1 | ERROR: [3] bootstrap checks failed |
A:
deploy环境准备
由于Elasticsearch、Logstash、Kibana均不能以root账号运行。
但是Linux对非root账号可并发操作的文件、线程都有限制。
所以，部署deploy相关的机器都要调整：
修改文件限制
修改系统文件
vi /etc/security/limits.conf
增加的内容
1 | soft nofile 65536 |
调整进程数：
修改系统文件
vi /etc/security/limits.d/20-nproc.conf
调整成以下配置
1 | soft nproc 4096 |
调整虚拟内存&最大并发连接：
修改系统文件
vi /etc/sysctl.conf
增加的内容，执行 “sudo sysctl -p” 生效：
1 | vm.max_map_count=655360 |
以上操作重启系统后生效: reboot
=======================================================
Q:
1 | fatal: [10.0.0.12]: FAILED! => {"changed": false, "msg": "Failed to find handler for "/root/.ansible/tmp/ansible-tmp-1530157101.0-36263796351235/source". Make sure the required command to extract the file is installed. Command "unzip" not found. Command "/usr/bin/gtar" could not handle archive."} |
A: 目标主机缺少unzip命令。
Solution：在远程目标主机上执行安装命令：
sudo yum install unzip
ansible remote_servers -i \_etc_ansible_hosts -m yum -a "list=unzip"