I do not see recovery as such but the logs are created and if following a
problem I re-start an operation from my web test programme anything to do
with persistence hangs until I restart the server having deleted the logs.
Also I have looked at the hung thread in a debugger and sees that it appears
to be hanging on a call from the the TransactionManager to recover. This is
all a bit vague I know, but the problem is intermittent and as would luck
would have it I can't reproduce it.
Should distributed transactions across two non -clusted server instances on
different machines on the network work? Or do I need to set up clustering -
we are trying to ascertain possible problems when we use Pre-Request load
blancing of stateless Session beans and are mimicking calls across boxes
using explicit EJB resources
e.g. @EJB(mappedName="corbaname:iiop:othermachine:3700#BeanX")?
--
[Message sent by forum member 'JFK9']
View Post: http://forums.java.net/node/786847