[keycloak-user] KeyCloak Clustering and High Availability question

[Bill Burke]

Wildfly clustering handles user login session replication as well as the 
caching layer for information stored in the shared database and/or LDAP 
and/or any other external user store.  Shared database is required for 
storing realm configuration data and user data and imported user data.  
So Wildfly clustering handles anything in-memory related and for 
database stuff you have to rely on your database vendor for whatever 
solution they have for this.


On 7/11/17 8:48 AM, Reza Shams Amiri wrote:
> Hi,
>
> I am also evaluating KeyCloak for my organization. I have a question about how failover in KeyCloak works.
>  From what I understood from the documentation, it says that the application scalability is handled by wildfly clustering but with a shared database.
> I couldn’t find a documentation about what we should do in case of database failure?
> We want to have two different clustered nodes in two different continents for idp and we mainly have mysql databases. Clustering them is actually painful and done through rabbitMQ synced messages and in some custom ways. So how can we handle database failure in KeyCloak let’s say if the link between Sweden and USA is completely broken?
>
> Thanks a lot
> /Reza
>
> _______________________________________________
> keycloak-user mailing list
> keycloak-user at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/keycloak-user