Hi,
I'm using GF 3.1.1 in a clustered deployment, I'm also using EJB Timer
Service, I've setup an external (not-Derby) database and pointed
jdbc/__TimerPool to it... everything is working as expected.
Doing some crash recovery tests I found that:
1) If glassfish is unable to connect to the TimerPool database, the timer
service stops executing scheduled jobs (That's ok, no DB, no jobs)
2) If connection is available again after a short period of time (2 mins) the
timer service goes back to normal and starts executing jobs again (as I
expected)
3) If connection is available again after a longer period of time (say 15
mins) the timer service does not goes back to normal and jobs are not
executed anymore (unexpected!)
Is it supposed to work this way?
Is there something wrong with my setup?
PS: connection validation is enabled in TimerPool connection pool.
--
[Message sent by forum member 'galmeida']
View Post: http://forums.java.net/node/855044