Re: ActiveMQ database lock problem

[mickhayes <mi...@gmail.com>]

I think it's best if you share as little as possible between the two brokers.
In this case, you are trying to share the port (in time). 

As an alternative to your configuration, could you use a different port for
each broker, and the failover transport on the client side?

Would that be good enough for your case?

If however, you still wanted to delay the startup of the second broker, I
would say that the only satisfactory end condition for the delay would be
that the socket is in state CLOSED. 


-----
Michael Hayes B.Sc. (NUI), M.Sc. (DCU), SCSA SCNA 

--
View this message in context: http://activemq.2283324.n4.nabble.com/ActiveMQ-database-lock-problem-tp4652786p4652791.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.

Re: ActiveMQ database lock problem

[mickhayes <mi...@gmail.com>]

Looks interesting... you might get a response if you comment on his blog
entry directly.

However, I think http://activemq.apache.org/jdbc-master-slave.html is worth
a look anyway:

"Clients should be using the Failover Transport to connect to the available
brokers. e.g. using a URL something like the following
failover:(tcp://broker1:61616,tcp://broker2:61616,tcp://broker3:61616)"

In the above case, the brokers are (most likely) on different machines - at
least that would be the intention when different machine names are given in
a example.

Your brokers are, however, on the same machine, so why not a different port
per broker instance?

Unless you have a very good reason to use the same port for each broker's
transport, I think you should use different ports.





-----
Michael Hayes B.Sc. (NUI), M.Sc. (DCU), SCSA SCNA 

--
View this message in context: http://activemq.2283324.n4.nabble.com/ActiveMQ-database-lock-problem-tp4652786p4652829.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.