tribe.asciidoc 2.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
  1. [[modules-tribe]]
  2. == Tribe node
  3. The _tribes_ feature allows a _tribe node_ to act as a federated client across
  4. multiple clusters.
  5. WARNING: This feature is EXPERIMENTAL -- use at your own risk.
  6. The tribe node works by retrieving the cluster state from all connected
  7. clusters and merging them into a global cluster state. With this information
  8. at hand, it is able to perform read and write operations against the nodes in
  9. all clusters as if they were local.
  10. The `elasticsearch.yml` config file for a tribe node just needs to list the
  11. clusters that should be joined, for instance:
  12. [source,yaml]
  13. --------------------------------
  14. tribe:
  15. t1: <1>
  16. cluster.name: cluster_one
  17. t2: <1>
  18. cluster.name: cluster_two
  19. --------------------------------
  20. <1> `t1` and `t2` are aribitrary names representing the connection to each
  21. cluster.
  22. The example above configures connections to two clusters, name `t1` and `t2`
  23. respectively. The tribe node will create a <<modules-node,node client>> to
  24. connect each cluster using <<multicast,multicast discovery>> by default. Any
  25. other settings for the connection can be configured under `tribe.{name}`, just
  26. like the `cluster.name` in the example.
  27. The merged global cluster state means that almost all operations work in the
  28. same way as a single cluster: distributed search, suggest, percolation,
  29. indexing, etc.
  30. However, there are a few exceptions:
  31. * The merged view cannot handle indices with the same name in multiple
  32. clusters. It will pick one of them and discard the other.
  33. * Master level read operations (eg <<cluster-state>>, <<cluster-health>>)
  34. need to have the `local` flag set to `true` as the tribe node does not
  35. have a single master node.
  36. * Master level write operations (eg <<indices-create-index>>) are not
  37. allowed. These should be performed on a single cluster.
  38. The tribe node can be configured to block all write operations and all
  39. metadata operations with:
  40. [source,yaml]
  41. --------------------------------
  42. tribe:
  43. blocks:
  44. write: true
  45. metadata: true
  46. --------------------------------