[keycloak-dev] HA mode with JDBC_PING shows warning in the logs after migration to 4.8.3 from 3.4.3

[abhishek raghav]

Thanks Sebastian.

I tried running the same setup with 5.0.0 of keycloak, I did not see any
such errors which I reported in my first email. This was definitely a
Wildfly issue and not keycloak.

Regarding my 2nd question - i.e. support of "clear_table_on_view_change"
property. I see that jgroups has removed support of this property. So lets
say if JGROUPSPING table has lot stale entries, while keycloak starts
booting up - each time keycloak node will try to JOIN with all the entries
already present in the JGROUPSPING table and thus time taken for the
service to start will be more. If that timeline is more than 300s, keycloak
does not start and reports timeout error.
This scenario is highly possible in cloud scenarios, since there the
keycloak nodes can start on any available host/IP since no of nodes are not
fixed.

Can you suggest any workaround to fix this.

*- Best Regards*
   Abhishek Raghav








On Fri, Apr 26, 2019 at 6:11 PM Sebastian Laskawiec <slaskawi at redhat.com>
wrote:

> There was a bunch of fixed to JGroups a while ago, including changes in
> JDBC_PING.
>
> Could you please rerun your setup with Keycloak >= 5.0.0? I believe some
> of the issues (or maybe even all of them) should be fixed.
>
> On Thu, Apr 25, 2019 at 7:19 PM abhishek raghav <abhi.raghav007 at gmail.com>
> wrote:
>
>> Hi
>>
>> After the migration of keycloak HA configurations from 3.4.3.Final to
>> 4.8.3.Final, I am seeing some WARNINGS on one of the nodes of keycloak
>> immediately after the keycloak is started with 2 nodes. This occurs after
>> every time when the cluster is scaled up or whenever infinispan is trying
>> to update the cluster member list.
>> I am using JDBC_PING to achieve clustering in keycloak.
>>
>> Below is the stacktrace -
>>
>> 2019-04-24 12:20:43,687 WARN
>> >> [org.infinispan.topology.ClusterTopologyManagerImpl]
>> >> (transport-thread--p18-t2) [dcidqdcosagent08] KEYCLOAK DEV 1.5.RC
>> >> ISPN000197: Error updating cluster member list:
>> >> org.infinispan.util.concurrent.TimeoutException: ISPN000476: Timed out
>> >> waiting for responses for request 1 from dcidqdcosagent02
>> >
>> >                   at
>> >>
>> org.infinispan.remoting.transport.impl.MultiTargetRequest.onTimeout(MultiTargetRequest.java:167)
>> >
>> >                   at
>> >>
>> org.infinispan.remoting.transport.AbstractRequest.call(AbstractRequest.java:87)
>> >
>> >                   at
>> >>
>> org.infinispan.remoting.transport.AbstractRequest.call(AbstractRequest.java:22)
>> >
>> >                   at
>> >> java.util.concurrent.FutureTask.run(FutureTask.java:266)
>> >
>> >                   at
>> >>
>> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
>> >
>> >                   at
>> >>
>> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
>> >
>> >                   at
>> >>
>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
>> >
>> >                   at
>> >>
>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
>> >
>> >                   at java.lang.Thread.run(Thread.java:748)
>> >
>> >                   Suppressed: org.infinispan.util.logging.TraceException
>> >
>> >                                     at
>> >>
>> org.infinispan.remoting.transport.Transport.invokeRemotely(Transport.java:75)
>> >
>> >                                     at
>> >>
>> org.infinispan.topology.ClusterTopologyManagerImpl.confirmMembersAvailable(ClusterTopologyManagerImpl.java:525)
>> >
>> >                                     at
>> >>
>> org.infinispan.topology.ClusterTopologyManagerImpl.updateCacheMembers(ClusterTopologyManagerImpl.java:508)
>> >
>> >
>>
>> Now after I searched, I really did not see anyone reported such error on
>> keycloak but there is similar bug reported in WILDLFY 14 and is
>> categorized
>> as a blocker in WILDLFY 14.This bug is already fixed in WILDLFY 15.
>> https://issues.jboss.org/browse/WFLY-10736?attachmentViewMode=list
>>
>> Now since keycloak 4.8 is also based on WILDLFY 14, these WARNINGS could
>> be
>> because of this blocker in WILDFLY 14.
>>
>> What should I do to get rid this error. Is this really a problem in
>> keycloak 4.8.3.Final. Did anyone notice any such issue while running
>> keycloak 4.8.3 in HA mode.
>> Is there a workaround to fix this.
>>
>>
>> One more thing we noticed is - It is regarding a property in JDBC_PING
>> protocol we are using in our 3.4.3 setup i.e. "clear_table_on_view_change"
>> but it is no more supported in 4.8 version. and thus the JGROUPSPING table
>> is filled up with lot of stale entries. Is there a workaround to clear the
>> table after view change in 4.8 also.
>>
>> Thanks
>> Abhishek
>> _______________________________________________
>> keycloak-dev mailing list
>> keycloak-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/keycloak-dev
>>
>