Run keycloak and postgresql with docker compose by two scenarios
- Keycloak with simple instance
- Keycloak with cluster (experimental for your understanding of how Keycloak clusters work)
This repo is mainly set for my blogs
Run Keycloak in docker with extenal DB
Run Keycloak locally with Docker compose
Keycloak Cluster with Docker Compose — Up and Running in Seconds
update /etc/hosts , add below lines
# keycloak
127.0.0.1 keycloak.com.au
On Windows, the file path is usually: c:\Windows\System32\Drivers\etc\hosts
Notes: In many online documents and videos, Keycloak experts often recommend starting the Keycloak service on localhost with a specific port. However, this practice is not advisable, especially when working in a real environment. Instead, it’s more practical to configure Keycloak with a DNS-ready setup. This also allows you to test HTTPS access with SSL certifications later on if needed.
git clone https://github.com/ozbillwang/keycloak-compose.git
cd keycloak-compose
docker compose up -d
docker ps -a
make sure all containers running well
access http://keycloak.com.au:8180
go with "Administration Console ", then login with admin / password
Yes, the solution is ready now with help from Niko Köbler (@dasniko) with his cool video
update /etc/hosts , add below lines
# keycloak
127.0.0.1 keycloak.com.au
On Windows, the file path is usually: c:\Windows\System32\Drivers\etc\hosts
git clone https://github.com/ozbillwang/keycloak-compose.git
cd keycloak-compose
docker compose -f docker-compose-cluster.yml up -d
- Check the health
docker ps -a
- Check logs with Cluster events
docker logs -f <kc1 or kc2 container id>
- Check the cluster logs, there should be two members in cluster pool now
Received new cluster view for channel ISPN: [b31f28d4c94a-31765|1] (2) [b31f28d4c94a-31765, bc873530c08b-24274] Starting rebalance with members [b31f28d4c94a-31765, bc873530c08b-24274]
access http://keycloak.com.au:8180
go with "Administration Console ", then login with admin / password
Test the fail over and cluster realiable.
- kill one keycloak container
docker ps -a
docker rm -f keycloak-compose-kc2
Check logs, you will only see one member in Cluster pool now.
Updating cache members list [b31f28d4c94a-31765], topology id 6
When you refresh the website http://keycloak.com.au:8180, it takes about 5~10 seconds at first time, then work as normal
- restore all services
$ docker compose -f docker-compose-cluster.yml up -d
✔ Container db Running
✔ Container kc1 Started # because I killed it before
✔ Container kc2 Running
✔ Container kc_lb Running
Check logs again, two members in cluster pool now.
If you access http://keycloak.com.au:8180, it works fine
Starting rebalance with members [b31f28d4c94a-31765, 462ae7fcf1a3-41736], phase READ_OLD_WRITE_ALL, topology id 7 Finished rebalance with members [b31f28d4c94a-31765, 462ae7fcf1a3-41736], topology id 10
https://www.keycloak.org/2019/05/keycloak-cluster-setup.html
https://www.youtube.com/watch?v=P96VQkBBNxU