I am completely lost with configuring the timer service in a clustered
environment, probably due to a lack of understanding. Although I am deploying
my ear into the cluster initially there is a lot of deployment activity in
the domains server.log, then it continues in the cluster's instance
server.log. If the timer service is not coming up for the domain it does not
continue to deploy to the cluster instance. So I concluded I have to enable
it for both. I already followed Oracle's hint of defining a custom jbdc
ressource for the timers instead of the default jdbc/__TimerPool. Also I
pointed the Timer Datasource setting of the EJB Timer service of both the
server config as well as the cluster config to my custom ressource. The
result I get is "Another instance of Derby may have already booted the
database." How to do that properly?
--
[Message sent by forum member 'ztangm']
View Post: http://forums.java.net/node/889555