Question: How Connect to a Neo4j Cluster? about bolt-rs HOT 2 CLOSED

I haven't implemented support for Neo4j clustering - that'll probably be a feature of the upcoming higher-level driver. If we assume a simple setup with 1 primary and several secondaries, and you want a way to get a connection pool to your primary server and then a list of connection pools for your read replicas, you could use the following approach, though it's probably not ideal:

1. Connect to your primary server and create a pool using the approach in the deadpool-bolt example. You can use this pool for reads and writes.
2. Grab a list of your read replicas by calling route on a connection from your primary pool. Use the list of IP addresses in the response to create a connection pool for each read replica, using the same approach as in step 1.
3. Distribute reads across your read replicas using whatever strategy you wish.

Note that I haven't tested this approach, and it doesn't account for things like multiple primaries or a routing table that changes over time - you'll have to handle those yourself if that's your use case. More details on clustering and the routing table here.

from bolt-rs.

Thanks for your detailed answer, I would try to implement it.

For example when I create the first pool and then get all the read replicas. I would store all these new pools in an Vec and the use something like a load balancer or round robin for each request.

I will be use a neo4j cluster o memgraph cluster

from bolt-rs.