activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stephan Burkard <>
Subject Expected state after failover
Date Fri, 30 Jan 2015 09:31:34 GMT

I would like to know what the expected state after a successful
master/slave failover is. The AMQ website seems to be inconsistent or at
least not very accurate in this topic.

The general master/slave page with the different flavours listed, says for
all types "Automatic recovery of old masters".

So I would expect that after a successful automatic failover, the former
master is a slave and the former slave (one of them if multiple) is master.
That means if other incidents occur, the automatic failover works again and
again without any manual intervention.

In contrast, the concrete page about shared-filesystem master/slave says
"If the master ... it immediately shuts down". And later "after a restart
of an old master".

This gives me the impression that after a successful failover, the former
master shuts down and the former slave becomes master. But I have to
restart the former master manually. In other words: when I have two
brokers, only one automatic failover can happen because after the first
incident, I still have a master broker but no slaves left.

Is my interpretation correct? And if yes, which one of the scenarios is
correct? Or is the expected state depending on the concrete master/slave


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message