Good morning ! @snak: That was the first hint I found during my investigation but it could not be the failure. It could not be a problem of version incompatibility. As I wrote the transaction was working 4 times and failed at the 5th transaction (reproducable). If it would be a problem of incompatibility no transaction could be successful. ConditionEvaluatorf4a3f354729241ac8370890200fdf2d8.evaluate(Unknown Source) Guess this class is generated automaticly and the source that is missing at this point is the kbase. I checked again the initialisation of kbase and I found a failure. The kbase is defined as static and is beeing initialised with the first transaction. Additionaly it's beeing updated every few minutes. After I corrected this the failure disappear. I did the following changes: *old code:* if ((knowledgeSetupTimestamp + timeInMillis) < System.currentTimeMillis()) ageReload = true; if ((kbase == null) || ageReload) { synchronized (DroolsHelper.class) { if ((kbase == null) || ageReload) { ......... } } } *new corrected code:* if ((kbase == null) || (knowledgeSetupTimestamp + timeInMillis < System.currentTimeMillis())) { synchronized (DroolsHelper.class) { if ((kbase == null) || (knowledgeSetupTimestamp + timeInMillis < System.currentTimeMillis())) { ......... } } } I'm still a bit puzzled about the reported failure and it would be great if someone could explain me this behaviour. Thanks a lot ! Cheers Johannes -- View this message in context: http://drools.46999.n3.nabble.com/java-lang-NoSuchMethodError-thrown-in-d... Sent from the Drools: User forum mailing list archive at Nabble.com.