Cross Cluster Search简介
cross-cluster search功能允许任何节点作为跨多个群集的federated client(联合客户端),与tribe node不同的是cross-cluster search节点并不会加入remote cluster(远程集群),而是用轻量的方法连接到remote cluster,以便执行federated search(联合搜索)
Remote cluster
要使用cross-cluster search之前需要先了解remote cluster
一个remote cluster中有"name"和seed nodes(种子节点)列表以供引用,注册remote cluster时,会从其中一个seed node来检查其集群状态,以便在默认情况下选择最多三个有资格的节点作为gateway nodes(网关节点), 集群中配置了remote cluster的每个节点都连接到一个或多个gateway nodes,并使用它们将federated search到remote cluster。
可以使用集群设置(可以动态更新)在全局指定remote cluster,也可以在各个节点中的elasticsearch.yml
指定remote cluster 。
如果节点通过elasticsearch.yml文件
配置remote cluster,则可以通过该节点连接到remote cluster。换句话说,federated search只有发送到该节点才能连接到remote cluster。通过cluster settings API 设置的remote cluster集群中的每个节点(设置了cluster.remote.connect: true的节点)都可以连接。
通过elasticsearch.yml设置
:
cluster: remote: cluster_one: seeds: 127.0.0.1:9300 cluster_two: seeds: 127.0.0.1:9301
cluster_one和cluster_two表示与每个群集连接的任意群集别名。这些名称之后用于区分本地和远程索引
使用cluster settings API设置:
PUT _cluster/settings { "persistent": { "cluster": { "remote": { "cluster_one": { "seeds": [ "127.0.0.1:9300" ] }, "cluster_two": { "seeds": [ "127.0.0.1:9301" ] }, "cluster_three": { "seeds": [ "127.0.0.1:9302" ] } } } } }
删除远程群集:
PUT _cluster/settings { "persistent": { "cluster": { "remote": { "cluster_three": { "seeds": null } } } } }
删除cluster_three保留cluster_one和cluster_tow
Remote cluster的设置:
cluster.remote.connections_per_cluster
gateway nodes数量,默认是3
cluster.remote.initial_connect_timeout
节点启动时等待远程节点的超时时间,默认是30s
cluster.remote.node.attr
一个节点属性,用于过滤掉remote cluster中 符合gateway nodes的节点,比如设置cluster.remote.node.attr=gateway,那么将匹配节点属性node.attr.gateway: true
cluster.remote.connect
默认情况下,群集中的任意节点都可以充当federated client并连接到remote cluster,cluster.remote.connect可以设置为 false(默认为true)以防止某些节点连接到remote cluster
cluster.remote.${cluster_alias}.skip_unavailable
在节点中跳过特定的群集别名,默认是false
使用cross-cluster search查询
要搜索远程集群cluster_one上的twitter索引,index名和集群别用冒号分开:
GET /cluster_one:twitter/_search { "query": { "match": { "user": "kimchy" } } }
与tribe特征相反,cross-cluster search还可以在不同群集上搜索相同名称的index:
GET /cluster_one:twitter,twitter/_search { "query": { "match": { "user": "kimchy" } } }
搜索结果的歧义与索引在请求中消除歧义的方式相同。即使index名称相同,这些index也会在合并结果时被视为不同的index。从远程index检索的所有结果都将以remote cluster的name为前缀:
{ "took": 150, "timed_out": false, "_shards": { "total": 2, "successful": 2, "failed": 0, "skipped": 0 }, "_clusters": { "total": 2, "successful": 2, "skipped": 0 }, "hits": { "total": 2, "max_score": 1, "hits": [ { "_index": "cluster_one:twitter", "_type": "_doc", "_id": "0", "_score": 1, "_source": { "user": "kimchy", "date": "2009-11-15T14:12:12", "message": "trying out Elasticsearch", "likes": 0 } }, { "_index": "twitter", "_type": "_doc", "_id": "0", "_score": 2, "_source": { "user": "kimchy", "date": "2009-11-15T14:12:12", "message": "trying out Elasticsearch", "likes": 0 } } ] } }
跳过已经断开连接的集群:
默认情况下,在执行搜索请求时,通过cross-cluster search搜索的所有remote cluster都必须可用,否则整个请求将失败,并且尽管某些群集可用,但不会返回搜索结果。可以通过skip_unavailable设置使remote cluster可选,默认设置为false。
PUT _cluster/settings { "persistent": { "cluster.remote.cluster_two.skip_unavailable": true } }
cluster_two就变成可选的了
GET /cluster_one:twitter,cluster_two:twitter,twitter/_search { "query": { "match": { "user": "kimchy" } } }
在本地、cluster_one
,cluster_two中
搜索索引twitter
{ "took": 150, "timed_out": false, "_shards": { "total": 2, "successful": 2, "failed": 0, "skipped": 0 }, "_clusters": { #clusters部分表示一个群集不可用并被跳过 "total": 3, "successful": 2, "skipped": 1 }, "hits": { "total": 2, "max_score": 1, "hits": [ { "_index": "cluster_one:twitter", "_type": "_doc", "_id": "0", "_score": 1, "_source": { "user": "kimchy", "date": "2009-11-15T14:12:12", "message": "trying out Elasticsearch", "likes": 0 } }, { "_index": "twitter", "_type": "_doc", "_id": "0", "_score": 2, "_source": { "user": "kimchy", "date": "2009-11-15T14:12:12", "message": "trying out Elasticsearch", "likes": 0 } } ] } }