Scala Play 2.2 Slick 1.0.1-未来{尝试{…}}等待免费可用连接超时

Question

我有一个scala.html页面，它在Play 2.2.1、Scala 2.10.2、Slick 1.0.1、Postgres 9.3应用程序中调用AJAX。

以下同步代码工作正常。它解析请求查询字符串并调用方法 Schools.findSchoolsByFilter，该方法对表进行 scala 光滑调用，并根据 SchoolFilter 对象过滤结果并返回 Try[List[School]]

def listSchools = Action { implicit request =>
  db.withSession { implicit s: Session =>
    Schools.findSchoolsByFilter(parseFilter) match {
      case Success(schools) => Ok(toJsArray(schools))
      case Failure(e) => Ok(e.getMessage)
    }
  }
}

如果我将方法改为异步运行(见下文)，并多次调用listSchools，那么大约20秒后就会抛出这个异常。我怀疑这可能是某种类似于后播放slick和Async的竞态条件——这是竞态条件吗？。我的问题是，我应该如何更改这段代码来安全地异步运行它？

def listSchools = Action.async { implicit request =>
  db.withSession { implicit s: Session =>
    Schools.findSchoolsByFilter(parseFilter) map {
      case Success(schools) => Ok(toJsArray(schools))
      case Failure(e) => Ok(e.getMessage)
    }
  }
}
def findSchoolsByFilter(f: SchoolFilter, n: Int)(implicit s: Session) = 
  future { Try {
    ...
}}
case class SchoolFilter(name: Option[String], 
                        city: Option[String], 
                        state: Option[String],
                        zip: Option[String], 
                        district: Option[String])

以下是我的依赖项：

libraryDependencies ++= Seq(
  jdbc,
  cache,
  "com.typesafe.slick" %% "slick" % "1.0.1",
  "com.github.tototoshi" %% "slick-joda-mapper" % "0.4.0",
  "org.scalatest" % "scalatest_2.10" % "2.0" % "test",
  "org.easymock" % "easymock" % "3.2",
  "org.postgresql" % "postgresql" % "9.3-1100-jdbc4"
)

以下是堆栈跟踪：

com . jol box . bone CP . defaultconnectionstrategy . get connection internal(defaultconnectionstrategy . Java:88)com . jol box . bone CP . abstractconnectionstrategy . get connection(abstractconnectionstrategy . Java:90)com . jol box . bone CP . bone CP . get connection(bone CP . Java:553)com . jol box . bone CP . bonecp data source . get connection(bone CP data source . Java:131)Scala . slick . session . database $ $ anon $ 1 . create connection(database . Scala:82)apply(try . Scala:161)models . school . schools $ $ anon fun $ findSchoolsByFilter $ 1 . apply(school . Scala:84)models . schools $ $ anon fun $ findSchoolsByFilter $ 1 . apply(school . Scala:84)Scala . concurrent . impl . future $ promisecompletingrunnable . lifted tree 1 $ 1(future . Scala:24)Scala . concurrent . impl . future $ promisecompletingrunnable . run(future . Scala:24

根据Heroku/Play/BoneCp连接问题的建议，我将“com.jolbox”%“BoneCp”%”0.8.0.RELEASE“添加到我的库依赖项中，并将以下内容添加到application中。conf＜/code＞，我仍然得到相同的行为。

db.default.idleMaxAge=10 minutes
db.default.idleConnectionTestPeriod=30 seconds
db.default.connectionTimeout=20 second
db.default.connectionTestStatement="SELECT 1"
db.default.maxConnectionAge=30 minutes




Answer 1

您需要在将来打开会话。现在你正在做相反的事情。光滑会话在 withSession 块结束后无效。如果您在块中打开未来，即使在 withSession 块结束并且会话无效之后，它也会挂起会话。如果将来的代码尝试使用无效会话，则最终会出现不可预测的行为。

或者换句话说，移动< code > db . with Session { implicit s:Session =