Miroslav Novak created ARTEMIS-665:
--------------------------------------
Summary: Optimize master/slave replication after failback - slave
does not have to resync journal from master after failback
Key: ARTEMIS-665
URL: https://issues.apache.org/jira/browse/ARTEMIS-665
Project: ActiveMQ Artemis
Issue Type: Improvement
Components: Broker
Affects Versions: 1.3.0
Reporter: Miroslav Novak
If there is master-slave pair with replicated journal and master is killed then
slave activates and all clients failover to slave.
If fail-back is enabled and master is started again then master will copy
journal from slave before it will activate and failback happens. So far good.
Problem is that once master activates then slave restart itself and starts to
replicate the whole journal from master again. Throwing away it's original
journal which just replicated to master. This is inefficiency and takes network
resources.
The improvement should be that slave will not throw away it's "up-to-date"
journal and once master activates, it will just continue to replicate from
master.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
