I am trying to setup keycloak cluster with a shared database in AWS environment. I followed all steps mentioned on this link: https://www.keycloak.org/docs/latest/server_installation/#_standalone-ha-mode Keycloak nodes are two AWS EC2 instances and shared database is AWS RDS. How does the keycloak instances identify each other in cluster in AWS (how does multicast work in a AWS VPC)? Because I’m constantly getting “no members discovered” error after running keycloak on both servers in HA configuration. Has anyone tried this before? Any references or things to take care of list would be great.
1 Like
I am also facing same issue.Can anyone help us out here will appreciate it.
1 Like
Use JDBC ping, as multicast is not supported in AWS EC2
https://developer.jboss.org/docs/DOC-16351
2 Likes
@ludimax: It worked! Thanks alot. 
can you please share your sample setup? Are you using Kubernetes?
Hello. I am also interested to use JDBC_PING in AWS rather that dealing with S3 buckets. We are running Keycloak in EC2s and we have a shared RDS MariaDB. If anyone can share any links or info it will be great and I thank you in advance!
Since its version 17, Keycloak migrated to Quarkus, which is better, but does some things differently.
You may benefit from this thread: Use of JDBC_PING with Keycloak 17 (Quarkus distro)
1 Like