ES配置生命周期策略

转载互联网

 

es可以用来存储日志,一般日志存储只是短期保存,超过一定时间日志要是能自动删除最好,这样保证索引文档不会过多,查询时效性也能得到保证。

本文参考的官网地址是:https://www.elastic.co/guide/en/elasticsearch/reference/7.6/getting-started-index-lifecycle-management.html  。

索引的生命周期分为四个阶段:HOT->WARM->COLD->DELETE。

上面除了HOT为必须的阶段外,其他为非必须阶段,可以任意选择配置。因为日志索引只要满足自己删除功能,所以下文只配置了HOT与DELETE阶段。

三步实现完成es生命周期管理

配置策略(policy)、索引模版(template)、索引(index)

 

1,配置策略(policy)

PUT _ilm/policy/datastream_policy 
{
  "policy": {                       
    "phases": {
      "hot": {                      
        "actions": {
              "rollover":{
                  "max_docs":1
              }
        }
      },
      "delete": {
        "min_age": "30s",           
        "actions": {
          "delete": {}              
        }
      }
    }
  }
}

创建了一个 “datastream_policy”策略,策略含义是:当hot阶段索引文档超过一条,就会发生rollover——便于测试,

delete阶段,超过30s执行删除动作。

还可以选择其它维度控制,索引大小,索引时间:

"max_size": "50GB"
"max_age": "30d"

2,索引模版(template)

PUT _template/datastream_template
{
  "index_patterns": ["datastream-*"],                 
  "settings": {
    "number_of_shards": 1,
    "number_of_replicas": 1,
    "index.lifecycle.name": "datastream_policy",      
    "index.lifecycle.rollover_alias": "datastream"    
  }
}

创建了一个索引模板,"index_patterns": ["datastream-*"],索引以datastream-开头的则自动采用setting的配置;

"index.lifecycle.name": "datastream_policy", 采用”datastream_policy“的策略;
"index.lifecycle.rollover_alias": "datastream",创建使用该模版创建的索引统一用”datastream“的别名进行管理

3,索引(index)

PUT datastream-000001
{
  "aliases": {
    "datastream": {
      "is_write_index": true
    }
  }
}

创建一个初始索引,并设置索引可写入。

 

配置完成,下面进行验证过程。

4,配置lifecycle检测时间

PUT /_cluster/settings
{ "transient": { "indices.lifecycle.poll_interval": "10s" } }

默认十分钟,便于测试,改为10s。

5,验证

POST datastream/_doc
{
"message":"hello world"
}

通常对外暴露别名,所以索引采用别名写入数据,
因为,别名”datastream“匹配index_patterns,写入后,es根据配置的策略“datastream_policy”进行rollover。
用以下命令查看索引rollover进度。

GET datastream-*/_ilm/explain

等待30秒后,返回的数据如下:

{
  "indices" : {
    "datastream-000002" : {
      "index" : "datastream-000002",
      "managed" : true,
      "policy" : "datastream_policy",
      "lifecycle_date_millis" : 1596442291089,
      "age" : "43.77s",
      "phase" : "hot",
      "phase_time_millis" : 1596442292420,
      "action" : "rollover",
      "action_time_millis" : 1596442301555,
      "step" : "check-rollover-ready",
      "step_time_millis" : 1596442301555,
      "phase_execution" : {
        "policy" : "datastream_policy",
        "phase_definition" : {
          "min_age" : "0ms",
          "actions" : {
            "rollover" : {
              "max_docs" : 1
            }
          }
        },
        "version" : 1,
        "modified_date_in_millis" : 1596441845846
      }
    },
    "datastream-000001" : {
      "index" : "datastream-000001",
      "managed" : true,
      "policy" : "datastream_policy",
      "lifecycle_date_millis" : 1596442291025,
      "age" : "43.83s",
      "phase" : "delete",
      "phase_time_millis" : 1596442330859,
      "action" : "delete",
      "action_time_millis" : 1596442330859,
      "step" : "wait-for-shard-history-leases",
      "step_time_millis" : 1596442330859,
      "phase_execution" : {
        "policy" : "datastream_policy",
        "phase_definition" : {
          "min_age" : "30s",
          "actions" : {
            "delete" : { }
          }
        },
        "version" : 1,
        "modified_date_in_millis" : 1596441845846
      }
    }
  }
}

索引自动创建了一个新的"datastream-000002",旧的索引”datastream-000001“ action变成DELETE。

再过一段时间再次查看索引

GET datastream-*/_ilm/explain

结果如下:

{
  "indices" : {
    "datastream-000002" : {
      "index" : "datastream-000002",
      "managed" : true,
      "policy" : "datastream_policy",
      "lifecycle_date_millis" : 1596442291089,
      "age" : "1.45m",
      "phase" : "hot",
      "phase_time_millis" : 1596442292420,
      "action" : "rollover",
      "action_time_millis" : 1596442301555,
      "step" : "check-rollover-ready",
      "step_time_millis" : 1596442301555,
      "phase_execution" : {
        "policy" : "datastream_policy",
        "phase_definition" : {
          "min_age" : "0ms",
          "actions" : {
            "rollover" : {
              "max_docs" : 1
            }
          }
        },
        "version" : 1,
        "modified_date_in_millis" : 1596441845846
      }
    }
  }
}

旧的索引已经删除,在做日志查询时,我们只需要使用别名”datastream“进行日志查询即可,es便可无感知完成日志存储删除动作。

 

posted @ 2021-05-13 16:01  技术颜良  阅读(1521)  评论(0编辑  收藏  举报