How to run a GDS write operation on a cluster and avoid errors "Writing results to the cluster is only available with a Licensed Neo4j Graph Data Science library deployed on a `Read Replica` instance"

Jun 25, 2024
Knowledge

Details

For running GDS write operations on Neo4j 4.4.x and 5.x Clusters, we have to meet the following requirements:

1. Cluster members

For Neo4j 4.4.x, the cluster must contain at least one Read Replica instance.

Single Core member and a Read Replica is a valid scenario.

For Neo4j 5.x, the cluster must contain at least one Secondary machine.

Single Primary and a Secondary is a valid scenario.

2. Enable Server Side Routing in cluster

On Neo4j 4.4.x, we make use of a Read Replica instance to deploy the GDS library and process analytical workloads. Calls to 'GDS write procedures' are internally directed to the cluster LEADER instance via server-side routing.

On Neo4j 5.x, calls to 'GDS write procedures' are internally directed via server-side routing to the cluster instance that is a Writer for the database we work on.

Neo4j 5.x supports different databases on the same machine to act as Primary or Secondary members of the cluster; we do not recommend mixed setup. GDS should be installed on a machine that is not serving transactional loads and does not participate in Leader elections.

3. GDS plugin deployment

On Neo4j 4.4.x

The GDS plugin is deployed on the Read Replica.

A valid GDS Enterprise Edition license must be installed and configured on the Read Replica.

The driver connection to operated GDS should be made using the bolt:// protocol, or server-policy routed to the Read Replica instance.

On Neo4j 5.x

The GDS plugin is deployed on the Secondary.

A valid GDS Enterprise Edition license must be installed and configured on the Secondary.

The driver connection to operated GDS should be made using the bolt:// protocol to the Secondary instance. We have to project the graph on the Read Replica or Secondary server as the same machine we run GDS write operation.

Note: We should use bolt:// protocol on both 4.4.x and 5.x when running GDS project and write operation.

Examples settings:

Below are the settings for Neo4j 4.4.15 one CORE and one Read Replica cluster:

Related part in CORE server neo4j.conf

dbms.mode=SINGLE
dbms.clustering.enable=true
dbms.routing.listenaddress=core:7618
dbms.routing.advertisedaddress=core:7618
dbms.routing.enabled=true
dbms.routing.default_router=SERVER
causal_clustering.discovery_listen_address=core:5000
causal_clustering.discovery_advertised_address=core:5000
causal_clustering.transaction_listen_address=core:6000
causal_clustering.raft_listen_address=core:7000
causal_clustering.raft_advertised_address=core:7000
causal_clustering.transaction_advertised_address=core:6000
gds.enterprise.license_file=/path_to_license_file/licenses/gds.key

Related part in Read Replica neo4j.conf

dbms.mode=READ_REPLICA
dbms.routing.listen_address=readrep:7628
dbms.routing.advertised_address=readrep:7628
dbms.routing.enabled=true
dbms.routing.default_router=SERVER
causal_clustering.initial_discovery_members=readrep:5000
causal_clustering.discovery_listen_address=readrep:5000
causal_clustering.discovery_advertised_address=readrep:5000
causal_clustering.transaction_listen_address=readrep:6000
causal_clustering.raft_listen_address=lab1.example.com:7000
causal_clustering.raft_advertised_address=readrep:7000
causal_clustering.transaction_advertised_address=readrep:6000
gds.enterprise.license_file=/path_to_license_file/licenses/gds.key

Below are the settings for Neo4j 5.4.0 one PRIMARY and one SECONDARY cluster:

Related part in PRIMARY neo4j.conf

initial.dbms.default_primaries_count=1
dbms.cluster.minimum_initial_system_primaries_count=1
initial.server.mode_constraint=PRIMARY
server.cluster.system_database_mode=PRIMARY
dbms.routing.enabled=true
dbms.routing.default_router=SERVER
gds.enterprise.license_file=/path_to_license_file/licenses/gds.key
server.routing.listen_address=:7618
server.routing.advertised_address=:7618
gds.enterprise.license_file=/path_to_license_file/licenses/gds.key

Related part in SECONDARY neo4j.conf

initial.dbms.default_primaries_count=1
dbms.cluster.minimum_initial_system_primaries_count=1
initial.server.mode_constraint=SECONDARY
server.cluster.system_database_mode=SECONDARY
dbms.routing.enabled=true
dbms.routing.default_router=SERVER
gds.enterprise.license_file=/path_to_license_file/licenses/gds.key
server.routing.listen_address=:7618
server.routing.advertised_address=:7618
gds.enterprise.license_file=/path_to_license_file/licenses/gds.key

On Neo4j 5.x versions, we should explicitly set the SECONDARY machine for the system database with the setting `server.cluster.system_database_mode=SECONDARY`, which means that this instance holds only a secondary copy of the system database.

Reference: https://neo4j.com/docs/graph-data-science/current/production-deployment/neo4j-cluster/

Title

How to run a GDS write operation on a cluster and avoid errors "Writing results to the cluster is only available with a Licensed Neo4j Graph Data Science library deployed on a `Read Replica` instance"

Article Created Date

3/15/2024, 10:55 PM

URL Name

15894486465299-How-to-run-a-GDS-write-operation-on-cluster-and-avoid-errors-Writing-results-to-the-cluster-is-only-available-with-a-Licensed-Neo4j-Graph-Data-Science-library-deployed-on-a-Read-Replica-instance

Last Modified Date

6/25/2024, 1:59 PM