爲什麼不用groovy和SqlException的事務回滾？

使用下面的爲什麼不用groovy和SqlException的事務回滾？

import groovy.sql.Sql 
def sql = Sql.newInstance("jdbc:mysql://localhost:3306/eventDb", "user", 
     "pwd", "com.mysql.jdbc.Driver") 

def sqlInsert = "INSERT INTO GTEST (EVENTID,TSTAMP,USER_ID,USER_FIRST_NAME) VALUES (?,?,?,?)" 
def sqlParams = ['EVENTID':0, 'TSTAMP':'','USER_ID':'janew','USER_FIRST_NAME':'janewithaverylongnamesothatitdoesntfitwell'] 
sqlParams['TSTAMP'] = new Date() 

sql.withTransaction {stmt -> 
    def eventId = sql.executeInsert("INSERT INTO EVENTS (LOGID,TSTAMP) VALUES (2,CURRENT_TIMESTAMP)") 
    sqlParams['EVENTID'] = eventId [0][0] 
    sql.executeInsert(sqlInsert, sqlParams.values().toList()) 
}

第二個執行會失敗，但是第一次插入不回滾。我在這裏做錯了什麼？

來源

2011-07-18 mosgjig