Head's Up! These forums are read-only. All users and content have migrated. Please join us at community.neo4j.com.
06-01-2020 12:20 AM
Hi
I have 6 node multi-dataCentre cluster, all are CORE nodes.
3 nodes are in US(called DC site) and 3 nodes are in ASIA (called DR site).
I have given following in neo4j.conf file :
dbms.mode=CORE
causal_clustering.minimum_core_cluster_size_at_formation=3
causal_clustering.minimum_core_cluster_size_at_runtime=3
My requirement is leader should always be elected from DC site and only when DC goes down leader should be elected from DR site
and 3 node cluster should work with both read and write.
I have tried with causal_clustering.refuse_to_be_leader=true on DR site but it didn't help because i need to reconfigure DR when DC goes down.
Thanks
Danny
06-01-2020 11:58 PM
06-02-2020 05:05 AM
@danny.oberoi this can not be achieved without manual intervention. if for example you set refuse_to_be_leader
in the DR site then for example on startup of all 6 instances then the leader would not be in the DR site as expected. However if the primary DC goes down then there would be no way to re-elect a leader in the DR site. You would need to stop an instance in the DR site, remove its refuse_to_be_leader
and then restart it for it then to become leader
All the sessions of the conference are now available online