Elasticsearch 是优秀的文档数据库，在我们使用集群方式创建我们的文档数据时，需要根据集群node数量合理设置分片个数 从而提高数据查询、读取 效率；
下面是分片设置块
"settings": {
"number_of_shards": 12,#分片个数，在创建索引不指定时 默认为 5；
"number_of_replicas": 1 #数据副本，一般设置为1；
},
下面是一个创建索引并设置分片的例子:
curl -X PUT http://$1:9200/your_index_name/ -H ‘content-type: application/json‘ -d ‘{
"settings": {
"number_of_shards": 12,
"number_of_replicas": 1
},
"mappings": {
"sms_up": {
"dynamic_templates": [
{
"strings_as_keywords": {
"match_mapping_type": "string",
"mapping": {
"type": "keyword"
}
}
}
],
"properties": {
"account_id": {
"type": "long"
},
"date": {
"type": "date",
"format": "yyyy-MM-dd"
},
"is_push_sms_up": {
"type": "short"
},
"mobile": {
"type": "keyword"
},
"push_sms_up_time": {
"type": "date",
"format": "yyyy-MM-dd HH:mm:ss"
},
"request_id": {
"type": "keyword"
},
"request_time": {
"type": "date",
"format": "epoch_second"
},
"sms_account_primary": {
"type": "integer"
},
"sms_content": {
"type": "keyword"
},
"sms_exno": {
"type": "keyword"
},
"sms_facilitator_id": {
"type": "integer"
},
"sms_msgid": {
"type": "keyword"
},
"sms_sign_id": {
"type": "integer"
},
"sms_spno": {
"type": "keyword"
},
"sms_type": {
"type": "long"
},
"sms_up_rtime": {
"type": "date",
"format": "yyyy-MM-dd HH:mm:ss"
}
}
}
}
}‘
原文：https://www.cnblogs.com/tank1992/p/10906840.html