If this was not clear: The bug in my program caused a exception on command.execute(jbpmContext) which I only could see by debugging the CommandServiceBean since the exception from jbpmContext.close() hid the original one. Corrrecting my proposal I'ld like to suggest that the exception from the finally block should only be ignored (i.e. only lead to an error message) if an exception has occured before. Volker View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4222475#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...