I've tried the masterslave discovery agent on broker amq3 & the failover from
master (amq1) to slave (amq2) works.
However, masterslave doesn't seem to support priorityBackup=true & so the
broker doesn't fail-back from slave to master when the master comes back
online.
My original issue is that I'm trying to implement fail-back & messages are
being lost after fail-back (failover works OK).
i.e.
- Broker amq3 priority-connects to amq1 (which bridges to amq2)
-> All messages delivered OK. (message route = 3->1->2)
- When broker amq1 is shutdown, broker amq3 fails-over to connect to amq2
-> All messages delivered OK (message route = 3->2)
- When broker amq1 is restarted, broker amq3 detects this & fails-back to
connect to amq1 instead of amq2
-> The amq1 broker doesn't seem to realise that this is a network
connection & so drops messages with duplicate producer-sequences even if
producerID is different. (message route = 3->1->2)
I'll update JIRA with my findings.
P.S.
My understanding of masterslave was that it was just a convenience wrapper
for
"static:(failover:(tcp://master,tcp://slave)?randomize=false&maxReconnectAttempts=0)"
--- does it do anything more than this under-the-hood?
ceposta wrote
> Looked briefly at your configs...
>
> Can you try this... for the brokers that have 2+, you can use the
> masterslave: discovery agent on your network connector... check out more
> details here: http://activemq.apache.org/networks-of-brokers.html
>
> For the NC that has single URI connectons, add this to the failover
> URI: ?&maxReconnectAttempts=0
>
> If that doesn't help, update the JIRA with this info and we can take a
> look..
--
View this message in context:
http://activemq.2283324.n4.nabble.com/Need-help-Messages-lost-when-using-failover-network-connection-with-failback-tp4671409p4671443.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
