-
Bug
-
Resolution: Done
-
Major
-
None
-
False
-
False
-
-
-
-
-
-
+
-
Undefined
-
In cluster environment when the ejb timers based on database store happens if the refresh interval is not small enough this scenario could happen:
Refreshing interval time (R)
Node 1 : creates an ejb timer at time T with expire date E.
Node 2 : tries to cancel an ejb timer within the interval [T, T+ R[
what happens is that any of the nodes could trigger at T + E the timer because the node 2 did not know the existance of the ejb timer node 1 created before the interval of refreshing.
This affects PAM product in cluster environment as we rely on ejb timers for timer features.
- is cloned by
-
WFLY-15342 EJB Timers out of sync within cluster causing inexistent timer to be triggered
- Closed
- is related to
-
RHPAM-4289 EJBTimerScheduler.getTimerByName inefficient with large number of timers and 'org.jbpm.ejb.timer.local.cache=false'
- Closed