You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by abhijith <to...@gmail.com> on 2017/05/26 15:32:30 UTC

Question on Artemis Shared store static config cluster

Hi,

I observed a behavior which was odd and would like to check if I am missing
something.  We have two node static configuration cluster.  One is master
and another one is slave with shared store.
---- master
 <ha-policy>
            <shared-store>
                <master>
                    <failover-on-shutdown>true</failover-on-shutdown>
                </master>
            </shared-store>
        </ha-policy>

---- slave
<ha-policy>
            <shared-store>
                <slave>
                    <failover-on-shutdown>true</failover-on-shutdown>
                    <allow-failback>false</allow-failback>
                </slave>
            </shared-store>
        </ha-policy>

This works fine as long as I bring up master.  We have tested killing master
server in which case slave takes over and then when master is bought back up
slave relinquishes control.  But, if we start slave up first before bringing
up master, then slave never starts servicing.  It sits there saying 

artemis-slave-local-out:11:38:26,267 INFO 
[org.apache.activemq.artemis.core.server] AMQ221020: Started NIO Acceptor at
myserver:61617 for protocols [CORE,MQTT,AMQP,HORNETQ,STOMP,OPENWIRE]
artemis-slave-local-out:11:38:26,268 INFO 
[org.apache.activemq.artemis.core.server] AMQ221010: Backup Server is now
live

If I start master and then kill it, then slave becomes active.  My
understanding of shared store HA cluster was that both master and slave are
trying to lock same file.  Whichever gets the lock file, they start
servicing.  Is my assumption incorrect?  Any clarification is really
appreciated.

Thanks
Abhi



--
View this message in context: http://activemq.2283324.n4.nabble.com/Question-on-Artemis-Shared-store-static-config-cluster-tp4726619.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.

Re: Question on Artemis Shared store static config cluster

Posted by Clebert Suconic <cl...@gmail.com>.
> This works fine as long as I bring up master.  We have tested killing master
> server in which case slave takes over and then when master is bought back up
> slave relinquishes control.  But, if we start slave up first before bringing
> up master, then slave never starts servicing.  It sits there saying
>


That is correct...

although at the first time the servers ever started, it will wait for
the live activation.

There's also the allowFailback, so whenever live starts, it would
shutdown itself.

I would recommend allowFailback = true with shared storage,

> artemis-slave-local-out:11:38:26,267 INFO
> [org.apache.activemq.artemis.core.server] AMQ221020: Started NIO Acceptor at
> myserver:61617 for protocols [CORE,MQTT,AMQP,HORNETQ,STOMP,OPENWIRE]
> artemis-slave-local-out:11:38:26,268 INFO
> [org.apache.activemq.artemis.core.server] AMQ221010: Backup Server is now
> live
>
> If I start master and then kill it, then slave becomes active.  My
> understanding of shared store HA cluster was that both master and slave are
> trying to lock same file.  Whichever gets the lock file, they start
> servicing.  Is my assumption incorrect?  Any clarification is really
> appreciated.
>
> Thanks
> Abhi
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.nabble.com/Question-on-Artemis-Shared-store-static-config-cluster-tp4726619.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.



-- 
Clebert Suconic