[keycloak-user] Clustering Keycloak via TCP

[lanabe]

I tested the following settings with 2 Keycloak nodes on the same
machine(using PostgreSQL), and it works.



---

<stack name="tcp">

  <transport type="TCP" socket-binding="jgroups-tcp"/>

  <protocol type="JDBC_PING">



    <property
name="datasource_jndi_name">java:jboss/datasources/KeycloakDS</property>

  </protocol>

  <protocol type="MERGE3"/>

  <protocol type="FD_SOCK" socket-binding="jgroups-tcp-fd"/>

  <protocol type="FD"/>

  <protocol type="VERIFY_SUSPECT"/>

  <protocol type="pbcast.NAKACK2"/>

  <protocol type="UNICAST3"/>

  <protocol type="pbcast.STABLE"/>

  <protocol type="pbcast.GMS"/>

  <protocol type="MFC"/>

  <protocol type="FRAG2"/>

</stack>

---

You can use TCPPING instead of JDBC_PING like this.

---

<protocol type="TCPPING">

  <property name="initial_hosts">

    <host1-ip-addr>[<jgroups-tcp port>],<host2-ip-addr>[<jgroups-tcp port>]

  </property>

</protocol>

---


On Tue, Feb 28, 2017 at 1:31 AM Bill Burke <bburke at redhat.com> wrote:

> Wildfly docs are wrong then :(
>
> Maybe this?  I'm just guessing.  You know I'm just googling stuff and
> looking at standalone-ha.xml, right?
>
> <subsystem xmlns="urn:jboss:domain:jgroups:4.0">
>      <channels default="ee">
>          <channel name="ee" stack="tcp"/>
>      </channels>
>
>
> On 2/27/17 10:59 AM, John D. Ament wrote:
> > That didn't work.
> >
> > Message: WFLYCTL0197: Unexpected attribute 'stack' encountered
> >
> > at org.jboss.as.controller.parsing.ParseUtils.unexpected
> >
> >
> >         <subsystem xmlns="urn:jboss:domain:infinispan:4.0">
> >
> >             <cache-container name="keycloak"
> > jndi-name="infinispan/Keycloak">
> >
> >                 <transport lock-timeout="60000" stack="tcp"/>
> >
> >
> > John
> >
> >
> > On Mon, Feb 27, 2017 at 10:30 AM Bill Burke <bburke at redhat.com
> > <mailto:bburke at redhat.com>> wrote:
> >
> >     https://docs.jboss.org/author/display/WFLY10/Infinispan+Subsystem
> >
> >     The only thing in keycloak that needs to leverage
> >     clustering/JGroups is
> >     our Infinispan cache for Keycloak SSO Sessions.  I've never
> personally
> >     tried to use a different JGRoups stack.  I believe you can set the
> >     "stack" attribute in the transport tag.  Let me know if you get it
> >     working.  I"ll expand on our docs.
> >
> >     <subsystem xmlns="urn:jboss:domain:infinispan:4.0">
> >          <cache-container name="keycloak"
> jndi-name="infinispan/Keycloak">
> >              <transport lock-timeout="60000"stack="tcp"/>
> >
> >
> >
> >     On 2/27/17 9:57 AM, John D. Ament wrote:
> >     > Hi
> >     >
> >     > I was wondering, has Keycloak been tested using Wildfly 10.1 TCP
> >     based
> >     > Jgroups?  I'm told that the TCP ports are lazy, and I'm never
> >     seeing them
> >     > come up.  It looks like Keycloak doesn't have a war file, no
> >     web.xml and as
> >     > a result no distributable flag.
> >     >
> >     > John
> >     > _______________________________________________
> >     > keycloak-user mailing list
> >     > keycloak-user at lists.jboss.org <mailto:
> keycloak-user at lists.jboss.org>
> >     > https://lists.jboss.org/mailman/listinfo/keycloak-user
> >
> >     _______________________________________________
> >     keycloak-user mailing list
> >     keycloak-user at lists.jboss.org <mailto:keycloak-user at lists.jboss.org>
> >     https://lists.jboss.org/mailman/listinfo/keycloak-user
> >
>
> _______________________________________________
> keycloak-user mailing list
> keycloak-user at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/keycloak-user
>