2017-01-05 10:10:34 -05:00
[[modules-cross-cluster-search]]
== Cross cluster search
The _cross cluster search_ feature allows any node to act as a federated client across
multiple clusters. In contrast to the _tribe_ feature, a _cross cluster search_ node won't
2017-01-09 12:17:43 -05:00
join the remote cluster, instead it connects to a remote cluster in a light fashion in order to execute
2017-01-05 10:10:34 -05:00
federated search requests.
The _cross cluster search_ feature works by configuring a remote cluster in the cluster state and connects only to a
limited number of nodes in the remote cluster. Each remote cluster is referenced by a name and a list of seed nodes.
Those seed nodes are used to discover other nodes eligible as so-called _gateway nodes_. Each node in a cluster that
2017-01-09 12:17:43 -05:00
has remote clusters configured connects to one or more _gateway nodes_ and uses them to federate search requests to
2017-01-05 10:10:34 -05:00
the remote cluster.
Remote clusters can either be configured as part of the `elasticsearch.yml` file or be dynamically updated via
the <<cluster settings API, cluster-update-settings>>. If a remote cluster is configured via `elasticsearch.yml` only
the nodes with the configuration set will be connecting to the remote cluster. Remote clusters set via the
<<cluster settings API, cluster-update-settings>> will be available on every node in the cluster.
The `elasticsearch.yml` config file for a _cross cluster search_ node just needs to list the
remote clusters that should be connected to, for instance:
[source,yaml]
--------------------------------
search:
remote:
seeds:
cluster_one: 127.0.0.1:9300 <1>
cluster_two: 127.0.0.1:9301 <1>
--------------------------------
<1> `cluster_one` and `cluster_two` are arbitrary names representing the connection to each cluster. These names are subsequently used to distinguish between local and remote indices.
[float]
=== Using cross cluster search
To search the `twitter` index on remote cluster `cluster_1` the index name must be prefixed with the cluster name
2017-01-05 11:03:12 -05:00
separated by a `:` character:
2017-01-05 10:10:34 -05:00
[source,js]
--------------------------------------------------
2017-01-05 11:03:12 -05:00
POST /cluster_one:twitter/tweet/_search
2017-01-05 10:10:34 -05:00
{
"match_all": {}
}
--------------------------------------------------
In contrast to the `tribe` feature cross cluster search can also search indices with the same name on different clusters:
[source,js]
--------------------------------------------------
2017-01-05 11:03:12 -05:00
POST /cluster_one:twitter,twitter/tweet/_search
2017-01-05 10:10:34 -05:00
{
"match_all": {}
}
--------------------------------------------------
Search results are disambiguated the same way as the indices are disambiguated in the request. Even if index names are
identical these indices will be treated as different indices when results are merged. All results retrieved from a remote index
2017-01-09 12:22:11 -05:00
will be prefixed with their remote cluster name:
2017-01-05 10:10:34 -05:00
[source,js]
--------------------------------------------------
{
"took" : 89,
"timed_out" : false,
"_shards" : {
"total" : 10,
"successful" : 10,
"failed" : 0
},
"hits" : {
"total" : 2,
"max_score" : 1.0,
"hits" : [
{
2017-01-05 11:03:12 -05:00
"_index" : "cluster_one:twitter",
2017-01-05 10:10:34 -05:00
"_type" : "tweet",
"_id" : "1",
"_score" : 1.0,
"_source" : {
"user" : "kimchy",
"postDate" : "2009-11-15T14:12:12",
"message" : "trying out Elasticsearch"
}
},
{
2017-01-05 11:03:12 -05:00
"_index" : "twitter",
2017-01-05 10:10:34 -05:00
"_type" : "tweet",
"_id" : "1",
"_score" : 1.0,
"_source" : {
"user" : "kimchy",
"postDate" : "2009-11-15T14:12:12",
"message" : "trying out Elasticsearch"
}
}
]
}
}
--------------------------------------------------
[float]
=== Cross cluster search settings
* `search.remote.connections_per_cluster` - the number of nodes to connect to per remote cluster. The default is `3`
* `search.remote.initial_connect_timeout` - the time to wait for remote connections to be established when the node starts. The default is `30s`.
* `search.remote.node_attribute` - a node attribute to filter out nodes that are eligible as a gateway node in the remote cluster.
For instance a node can have a node attribute `node.attr.gateway: true` such that only nodes with this attribute
will be connected to if `search.remote.node_attribute` is set to `gateway`
