Failover Postgresql Cluster
Hi,
I recently set up a PostgreSQL DB cluster using the Linode template and noticed that all nodes were configured as hot_standby, making them appear as primary nodes.
After reconfiguring node 1 to be a replica and verifying with repmgr, I now have one primary and two standby nodes.
My question is: which IP should I connect to? Currently, I'm using the private IP of Node1, configured to be accessible only within my internal network.
However, I'm unsure about what would happen in the event of a failover. As I understand it, I should be connecting through a load balancer, but the template didn't include one.
What's the correct approach here? :)
1 Reply
I think I found it…
https://www.linode.com/docs/products/compute/compute-instances/guides/failover/#ip-failover-support
can you confirm that I need to create a shared ip for those nodes and configure lelastic on each node bymyself?