elasticsearch设置共享目录、创建备份、恢复备份

假设es集群192.168.0.101、192.168.0.102、192.168.0.103三台节点

以下一至六步骤中除第三步只在101上执行外，其他步骤所有服务器都要执行

一、安装nfs服务

yum install nfs-utils rpcbind -y;

#设置开机自启动
systemctl enable rpcbind.service    
systemctl enable nfs-server.service

#启动服务
systemctl start rpcbind.service    
systemctl start nfs-server.service

二、创建共享目录

mkdir /data/elastic/bak/backup_es

# 由于备份程序是ES进程进行创建，因此设置目录的拥有者为启动ES程序的用户
chown -R es:es /data/elastic/bak/backup_es

三、为服务端修改配置文件

只在192.168.0.101上执行

vi /etc/exports

# 添加下面语句
/data/elastic/bak/backup_es *(rw,sync,no_root_squash,no_subtree_check)

# 刷新配置使得修改立刻生效
exportfs -a

# 查看可挂载目录
showmount -e 192.168.0.101

四、挂载目录

将三台都挂载到同一台服务器上

# 查看可挂载目录
showmount -e 192.168.0.101

# 挂载
mount -t nfs 192.168.0.101:/data/elastic/bak/backup_es /data/elastic/bak/backup_es

五、设置开机自动挂载

# 查看当前挂载
df -h

# 设置开机自动挂载
vi /etc/fstab
192.168。0.101:/data/elastic/bak/backup_es /data/elastic/bak/backup_es nfs defaults 0 0

六、修改es配置

vim elasticsearch.yml

# 在 elasticsearch.yml 中添加下面配置来设置备份仓库路径
path.repo: ["/data/elastic/bak/backup_es"]

七、重启es

八、创建备份仓库

POST _snapshot/ backup_es
{
  "type": "fs",
  "settings": {
    "location": "/data/elastic/bak/backup_es",
    "max_restore_bytes_per_sec": "100mb",
    "max_snapshot_bytes_per_sec": "100mb",
    "compress": true
  }
}

九、制作备份

POST _snapshot/ backup_es/ship_20191130  #_snapshot/备份仓库/备份名称
{
  "indices": "t_ship_latlng201911*",   #需要备份的索引
  "include_global_state": false,
  "ignore_unavailable": true
}

查看快照
GET _snapshot/backup_es/_all

十、恢复备份

# 全部恢复
POST /_snapshot/backup_es/ship_20191130/_restore

# 恢复指定的索引
POST /_snapshot/backup_es/ship_20191130/_restore
{
  "indices": "t_ship_latlng20191101,t_ship_latlng20191102",
  "ignore_unavailable": true,
  "include_global_state": false,
  "rename_pattern": "index_(.+)",
  "rename_replacement": "restored_index_$1",
  "index_settings": {
    "index.number_of_replicas": 0
  },
  "ignore_index_settings": [
    "index.refresh_interval"
  ]
}

代码解析详见：

https://blog.csdn.net/woshixiazaizhe/article/details/83896750