es 删除文档

创建索引：

PUT /my_alias_v3
{
"settings": {
"index": {
"number_of_shards": 2,
"number_of_replicas": 1
}
}
}

PUT my_alias_v2/docs/_mapping
{
  "properties": {
    "id": {"type": "long"},
    "name": {"type": "text"},
    "counter": {"type": "integer"},
    "tags": {"type": "text"}
  }
}

写入四个文档：

POST my_alias_v2/docs/_bulk
{"index": {"_id": 1}}
{"id":1, "name": "张三", "counter":"10", "tags":["red", "black"]}
{"index": {"_id": 2}}
{"id":2, "name": "李四", "counter":"20", "tags":["green", "purple"]}

POST my_alias_v2/docs/_bulk
{"index": {"_id": 3}}
{"id":3, "name": "haxi", "counter":"10", "tags":["red", "black"]}
{"index": {"_id": 4}}
{"id":4, "name": "xiha", "counter":"20", "tags":["green", "purple"]}

查询一个文档
GET my_alias_v2/_search
{
"query": {
"term": {
"id": {
"value": 4
}
}
}
}

删除一个文档
POST my_alias_v1/docs/_delete_by_query
{
"query": {
"term": {
"id": {
"value": 4
}
}
}
}

查看一个索引，删除后有一个delete的文档标记
GET _cat/shards/my_alias_v2?v

看segement都有deleted的标记文档
GET _cat/segments/my_alias_v2?v

强制合并索引
强制合并segement后主分片没有delete标记的文档了，但是副本分片多了一个segement。这时候主副本分片segement数量不一致，存储大小也是不一致的。原因是什么呢？
POST my_alias_v2/_forcemerge?max_num_segments=1

执行commit并清空translog的行为,确保所有数据都被写入磁盘
POST my_alias_v2/_flush

重新再执行一遍所有的流程，无论是一个主副分片还是两个主副分片，发现不会再出现强制合并索引大小不一致的情况了。搞不懂原因是什么。

elasticsearch中forcemerge清除文件占用的磁盘空间

https://blog.csdn.net/tiancityycf/article/details/115736887

猜你喜欢