使用Squeryl和Play!时遇到了一个奇怪的问题。
正常情况下,一切都完全正常。但是,如果在同一次请求中使用多个事务,就会出现错误。
这是我设置Squeryl的方式:
def initDB() {
import org.squeryl._
import play.db.DB
Class.forName("com.mysql.jdbc.Driver")
SessionFactory.concreteFactory = Some(() =>
Session.create( DB.getConnection, new MySQLAdapter) )
}
以下是样例事务,也是下面堆栈跟踪中引用的事务:
transaction {
import models.Game
Game.planets.insert(planetList)
Game.moons.insert(moonList)
}
堆栈跟踪:
Internal Server Error (500) for request GET /generate-galaxy
Execution exception (In /app/Generator.scala around line 330)
SQLException occured : You can't operate on a closed Connection!!!
play.exceptions.JavaExecutionException: You can't operate on a closed Connection!!!
at play.mvc.ActionInvoker.invoke(ActionInvoker.java:228)
at Invocation.HTTP Request(Play!)
Caused by: java.sql.SQLException: You can't operate on a closed Connection!!!
at com.mchange.v2.sql.SqlUtils.toSQLException(SqlUtils.java:106)
at com.mchange.v2.sql.SqlUtils.toSQLException(SqlUtils.java:65)
at org.squeryl.dsl.QueryDsl$class._executeTransactionWithin(QueryDsl.scala:95)
at org.squeryl.dsl.QueryDsl$class.transaction(QueryDsl.scala:64)
at org.squeryl.PrimitiveTypeMode$.transaction(PrimitiveTypeMode.scala:40)
at generator.Generator$$anonfun$generatePlanets$2.apply(Generator.scala:330)
at generator.Generator$$anonfun$generatePlanets$2.apply(Generator.scala:55)
at generator.Generator$.generatePlanets(Generator.scala:55)
at generator.Generator$.generateGalaxy(Generator.scala:36)
at controllers.MainRouter$.generateGalaxy(MainRouter.scala:29)
at play.mvc.ActionInvoker.invokeWithContinuation(ActionInvoker.java:543)
at play.mvc.ActionInvoker.invoke(ActionInvoker.java:499)
at play.mvc.ActionInvoker.invokeControllerMethod(ActionInvoker.java:493)
at play.mvc.ActionInvoker.invokeControllerMethod(ActionInvoker.java:470)
at play.mvc.ActionInvoker.invoke(ActionInvoker.java:158)
... 1 more
Caused by: java.lang.NullPointerException
... 14 more
我知道问题不在我的查询上,因为当使用scalatra作为web框架时,它们运行良好。我可以将所有内容放入一个事务块中,但那并不真正优雅,而且我也不确定在这种情况下是否会起作用 - planetList
列表有大约三百万成员,这使得Scala在将数据库插入分成50k个元素的较小块之前耗尽了内存。