Head's Up! These forums are read-only. All users and content have migrated. Please join us at community.neo4j.com.
05-14-2020 05:13 AM
hii guys ,
we have a neo4j enterprise edition running on 4 node cluster each having 16GegaBytes of RAM , there is one node with leader role and 3 other are follower role ,now i increased the heap size to 16GB to improve the performance and when i run the query it was using 100% of memory(16G) of leader node which mean it is not using any memory from follower nodes, and i don't have much idea about using other nodes memory to improve the performance.
i would be grateful if you can help me out on the same.
Thankyou
05-14-2020 01:15 PM
Transactions must be explicitly marked as being read or write transactions when executing them from your driver. You would typically use transactional functions to do this.
When this is in place, write transactions (when using a write transactional function) will be routed to the leader, and read transactions (when using read transactional functions) will be routed round-robin among the followers (and read replicas, if present).
If you are not using transactional functions, and not otherwise setting whether you're in read or write mode, then by default all transactions would be routed to the leader.
https://neo4j.com/docs/driver-manual/current/cypher-workflow/#driver-access-modes-transaction
All the sessions of the conference are now available online