tbain98 wrote
> In your broker-to-broker networkConnectors, are you using maxReconnects=0
> as an argument to the failover URI? It wouldn't explain why amq4 got
> demoted, but it could explain why messages aren't transferring to amq5
> instead.
Here is the definition of my networkConnectors inside activemq.xml:
<networkConnectors>
<networkConnector name="link1"
duplex="true"
conduitSubscriptions="false"
decreaseNetworkConsumerPriority="false"
uri="masterslave:(tcp://uromahn-amq4:61616,tcp://uromahn-amq5:61616,tcp://uromahn-amq6:61616)"/>
<networkConnector name="link2"
duplex="true"
conduitSubscriptions="false"
decreaseNetworkConsumerPriority="false"
uri="masterslave:(tcp://uromahn-amq4:61616,tcp://uromahn-amq5:61616,tcp://uromahn-amq6:61616)"/>
</networkConnectors>
As you can see, I have used the "masterslave" transport instead of
"failover" and have not specified any additional configuration parameters.
I will try, however, to change this to "failover" for another test.
tbain98 wrote
> You say you've got duplex connections between the clusters; which cluster
> is the one that establishes them via a networkConnector? And do you see
> the same behavior if you put producers on cluster2 and consumers on
> cluster1?
I have not tried that, but will do this first, even before I change
"masterslave" to "failover".
tbain98 wrote
> Also, looking at your logs it's not clear what happens between 13:00:48
> (when amq5 becomes the master) and 13:32:20 (30 minutes later, when the
> LevelDB exception occurs). Are messages transferring successfully to
> amq5,
> or is it sitting idle?
Oops. that was a copy&paste error. The message at 13:32:20 is actually
caused by me shutting down all my brokers. Sorry for the confusion.
And to answer your question: there were absolutely no messages transferred
from amq1 to either amq4, amq5, or amq6. This was visible by looking at the
"Network" page in the web console of amq1 which showed the number of
messages enqueued and dequeued.
I will post back to this thread once I could run the other suggested tests.
(removed my original post for brevity)
--
View this message in context:
http://activemq.2283324.n4.nabble.com/Potential-Bug-in-Master-Slave-with-Replicated-LevelDB-Store-tp4686450p4686488.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
