Data and processing distribution across nodes in Enterprise edition

I installed the Enterprise version on a cluster with 4 machines with high availability set to off. I have a couple of questions regarding how data and processing are distributed across the machines in the cluster:

  1. When I create a graph, is the graph data only stored in the leader node? Or the graph data are stored on every machine?

  2. When I run a query, will it automatically be distributed to all machines for processing?

Cheers,
Jim

  1. No. The graph data is partitioned across the cluster of nodes (every machine).

  2. If you use the “distributed” key word in your query, such as “create distributed query test()” , all the nodes of the cluster will participate the query processing.

I recommend this video to you which contains the architecture overview.

1 Like

Thanks @mingxiwu!

Some related questions on the distributed query: Is it possible to remove a node from the pool? Also is it possible the change the priority of certain nodes so that nodes with higher priority get assigned to a query first?