一. 引文:
部署了ELK 用来分析nginx的log,新增需求: 统计相关url(字段名url_path)访问次数最高的五个,通过Visualize创建走势图,发现自动对其进行了分词输出(如:/photo/pv/addPv被拆分成了photo,pv,addPV)。通过查询相关资料,明白了在创建索引,插入数据时,有一个template和mappings机制。
二. 概念:
template: 索引可使用预定义的模板进行创建,这个模板称作Index templates(索引模板)。模板设置包括settings和mappings,通过模式匹配的方式使得多个索引重用一个模板(如xxx* ,即:对xxx开头的index都匹配)。
mappings: 类似于静态语言中的数据类型:声明一个变量为int类型的变量, 以后这个变量都只能存储int类型的数据。同样的, 一个number类型的mapping字段只能存储number类型的数据。同语言的数据类型相比,mapping还有一些其他的含义,mapping不仅告诉ES一个field中是什么类型的值, 它还告诉ES如何索引数据以及数据是否能被搜索到。一个mapping由一个或多个analyzer组成, 一个analyzer又由一个或多个filter组成的。当ES索引文档的时候,它把字段中的内容传递给相应的analyzer,analyzer再传递给各自的filters。一个filter就是一个转换数据的方法, 输入一个字符串,这个方法返回另一个字符串(如一个将字符串转为小写的方法)。一个analyzer由一组顺序排列的filter组成,执行分析的过程就是按顺序一个filter一个filter依次调用, ES存储和索引最后得到的结果。总结来说, mapping的作用就是执行一系列的指令将输入的数据转成可搜索的索引项。
三. 核心:
elasticsearch默认是字符串的类型的字段都会分词(也会带来性能及空间占用问题)的(也就是引文中url_path问题所在)。故,这种情况下,解决办法就是修改mapping的string类型改为不分词("index": "not_analyzed")
需要注意的一个问题就是如果index中已经有大量数据的情况下,修改了analyzer,还需要清理已存在的数据,重新灌/cha入才能生效(因为就算put了新的mapping,以前的数据也更正不过来,后进去的数据会跟随旧数据的存储类型) 。。。
四. 基础操作:
可通过以下请求查看template及mapping:
# curl -XGET "http://10.4.30.41:9200/_template/?pretty" #结果太长,略
也可查看某一个template的:
#curl -XGET "http://10.4.30.41:9200/_template/logstash?pretty"
{
"logstash" : {
"order" : 0,
"template" : "logstash-*",
"settings" : {
"index" : {
"refresh_interval" : "5s"
}
},
"mappings" : {
"_default_" : {
"dynamic_templates" : [ {
"message_field" : {
"mapping" : {
"fielddata" : {
"format" : "disabled"
},
"index" : "analyzed",
"omit_norms" : true,
"type" : "string"
},
"match_mapping_type" : "string",
"match" : "message"
}
}, {
"string_fields" : {
"mapping" : {
"fielddata" : {
"format" : "disabled"
},
"index" : "analyzed",
"omit_norms" : true,
"type" : "string",
"fields" : {
"raw" : {
"ignore_above" : 256,
"index" : "not_analyzed",
"type" : "string"
}
}
},
"match_mapping_type" : "string",
"match" : "*"
}
} ],
"_all" : {
"omit_norms" : true,
"enabled" : true
},
"properties" : {
"@timestamp" : {
"type" : "date"
},
"geoip" : {
"dynamic" : true,
"properties" : {
"ip" : {
"type" : "ip"
},
"latitude" : {
"type" : "float"
},
"location" : {
"type" : "geo_point"
},
"longitude" : {
"type" : "float"
}
}
},
"@version" : {
"index" : "not_analyzed",
"type" : "string"
}
}
}
},
"aliases" : { }
}
}
可查看一个index的mapping:
#curl -XGET "http://10.4.30.41:9200/photo-xxx/_mapping/?pretty"
{
"photo-xxx" : {
"mappings" : {
"photo-xxx" : {
"properties" : {
"@timestamp" : {
"type" : "date",
"format" : "strict_date_optional_time||epoch_millis"
},
"@version" : {
"type" : "string"
},
"body_bytes_sent" : {
"type" : "string"
},
"bytes_sent" : {
"type" : "string"
},
"host" : {
"type" : "string"
},
"http_x_forwarded_for" : {
"type" : "string"
},
"message" : {
"type" : "string"
},
"method" : {
"type" : "string"
},
"path" : {
"type" : "string"
},
"remote_addr" : {
"type" : "string"
},
"request" : {
"type" : "string"
},
"request_time" : {
"type" : "string"
},
"servers" : {
"type" : "string"
},
"status" : {
"type" : "string"
},
"tags" : {
"type" : "string"
},
"time_local" : {
"type" : "string"
},
"type" : {
"type" : "string"
},
"upstream_addr" : {
"type" : "string"
},
"upstream_response_time" : {
"type" : "string"
},
"upstream_status" : {
"type" : "string"
},
"uri" : {
"type" : "string"
},
"url_args" : {
"type" : "string"
},
"url_path" : {
"type" : "string"
},
"verb" : {
"type" : "string"
}
}
}
}
}
}
解决分词问题:
新建一个template,如匹配"*xxx_*"的string都不分词,操作如下:
# curl -XPUT 'http://10.4.30.41:9200/_template/xxx?pretty' -d '{"template":"*xxx_*","mappings":{"_default_":{"dynamic_templates":[{"template_1":{"mapping":{"index": "not_analyzed","type": "string"},"match_mapping_type": "string","match": "*"}}],_all:{"enabled": false}}}}'
{
"acknowledged" : true
}
#注:其中match是匹配((unmatch非匹配))项(如本例中nginx的message,uri等)
至于数据这块,并未正式上线, 故可直接删除数据,生成新的即可(有空调研下相关导数据方法)。
再次查看mapping:
#curl -XGET "http://10.4.30.41:9200/photo-xxx/_mapping/?pretty"
{
"photo-xxx" : {
"mappings" : {
"_default_" : {
"_all" : {
"enabled" : false
},
"dynamic_templates" : [ {
"template_1" : {
"mapping" : {
"index" : "not_analyzed",
"type" : "string"
},
"match" : "*",
"match_mapping_type" : "string"
}
} ]
},
"photo-xxx" : {
"_all" : {
"enabled" : false
},
"dynamic_templates" : [ {
"template_1" : {
"mapping" : {
"index" : "not_analyzed",
"type" : "string"
},
"match" : "*",
"match_mapping_type" : "string"
}
} ],
"properties" : {
"@timestamp" : {
"type" : "date",
"format" : "strict_date_optional_time||epoch_millis"
},
"@version" : {
"type" : "string",
"index" : "not_analyzed"
},
"body_bytes_sent" : {
"type" : "string",
"index" : "not_analyzed"
},
"bytes_sent" : {
"type" : "string",
"index" : "not_analyzed"
},
"host" : {
"type" : "string",
"index" : "not_analyzed"
},
"http_x_forwarded_for" : {
"type" : "string",
"index" : "not_analyzed"
},
"message" : {
"type" : "string",
"index" : "not_analyzed"
},
"method" : {
"type" : "string",
"index" : "not_analyzed"
},
"path" : {
"type" : "string",
"index" : "not_analyzed"
},
"remote_addr" : {
"type" : "string",
"index" : "not_analyzed"
},
"request" : {
"type" : "string",
"index" : "not_analyzed"
},
"request_time" : {
"type" : "string",
"index" : "not_analyzed"
},
"servers" : {
"type" : "string",
"index" : "not_analyzed"
},
"status" : {
"type" : "string",
"index" : "not_analyzed"
},
"tags" : {
"type" : "string",
"index" : "not_analyzed"
},
"time_local" : {
"type" : "string",
"index" : "not_analyzed"
},
"type" : {
"type" : "string",
"index" : "not_analyzed"
},
"upstream_addr" : {
"type" : "string",
"index" : "not_analyzed"
},
"upstream_response_time" : {
"type" : "string",
"index" : "not_analyzed"
},
"upstream_status" : {
"type" : "string",
"index" : "not_analyzed"
},
"uri" : {
"type" : "string",
"index" : "not_analyzed"
},
"url_args" : {
"type" : "string",
"index" : "not_analyzed"
},
"url_path" : {
"type" : "string",
"index" : "not_analyzed"
},
"verb" : {
"type" : "string",
"index" : "not_analyzed"
}
}
}
}
}
}
入新数据后, 问题解决。。

如需要针对某些字段进行聚合统计/计算, 就需要设置字段类型为整数(integer)或浮点数(float)等。如在已创建kibana index的基础上修改字段类型, 需要删除kibana index,然后重新创建, 否则在kibana中不会生效(报错:No Compatible Fields: The "logstash-*" index pattern does not contain any of the following field types: number)。