Primary Shard and Replica Shard

Primary Shard and Replica Shard

Primary Shard

Primary Shard，即主本分片。每个文档都会存储在主本分片中。具体地，当索引一条文档（indexing a document）时，Elasticsearch首先将该文档保存在主本分片中，然后将其拷贝至主本分片的所有副本分片中。默认地，一个索引拥有一个主本分片；在创建索引时，我们可以通过number_of_shards参数指定主本分片的数量，如下：

PUT /<index-name>
{
    "settings": {
        "index": {
            "number_of_shards": 3
        }
    }
}

索引一旦生成，那么就不允许修改主本分片数量，因为文档归属于哪一个主本分片是由如下公式计算得出的，如果主本分片数量发生变更，那么就破坏了这种路由规则。

shard_num = hash(_routing) % num_primary_shards

Replica Shard

Replica Shard，即副本分片。副本分片是主本分片的拷贝，每个主本分片可以拥有零个或者多个副本分片，但主本分片与其副本分片不会落在同一Elasticsearch节点中。副本分片有以下优点：

• increase failover, a replica shard can be promoted to a primary shard if the primary fails.
• increase performance, get and search requests can be handled by primary or replica shards.

默认地，每个主本分片拥有一个副本分片；索引一旦生成，我们依然可以通过number_of_replicas参数动态地修改副本分片数量，如下：

PUT /<index-name>/_settings
{
    "settings": {
        "index": {
            "number_of_replicas": 2
        }
    }
}

A shard is a single Lucene instance. It is a low-level “worker” unit which is managed automatically by Elasticsearch.