[keycloak-dev] Issues with Infinispan local cache on EAP 6.4

[Stian Thorgersen]

----- Original Message -----
> From: "Libor Krzyžanek" <lkrzyzan at redhat.com>
> To: "Stian Thorgersen" <stian at redhat.com>
> Cc: "keycloak-dev" <keycloak-dev at lists.jboss.org>
> Sent: Wednesday, 2 September, 2015 4:05:44 PM
> Subject: Re: [keycloak-dev] Issues with Infinispan local cache on EAP 6.4
> 
> Ouch … question:
> 
> We have 2 nodes cluster to provide failover with user session replication.
> 
> Is this workaround needed? If yes, will user session be replicated to second
> node if first dies?

In your case the workaround is not activated as you're using clustering. It's only activated for local cache (no clustering).

The reason we need the workaround is that we dropped our old mem user session provider in favor of always using Infinispan. In non-clustered mode it uses a local cache, but that didn't work on EAP 6.4, so we had to put in this temporary workaround.

> 
> Thanks,
> 
> Libor Krzyžanek
> jboss.org <http://jboss.org/> Development Team
> 
> > On 02 Sep 2015, at 14:44, Stian Thorgersen <stian at redhat.com> wrote:
> > 
> > I recently removed the memory user session provider and replaced it with
> > Infinispan local caches. There's an issue with that on EAP though.
> > Infinispan didn't support map reduce tasks on local caches until 5.3 and
> > EAP 6.4 is on 5.2.
> > 
> > As a work around the Infinispan user session provider will fallback to the
> > old deprecated memory user session provider if Infinispan is older than
> > 5.3 and the cache is a local cache. The memory user session provider is
> > not available as a standalone provider, just used internally by the
> > Infinispan user session provider in this particular case.
> > 
> > Once we move to EAP 7 we can remove this work around.
> > _______________________________________________
> > keycloak-dev mailing list
> > keycloak-dev at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/keycloak-dev
> 
>