Akka的fault tolerant
要想容错，该怎么办？
父actor首先要获知子actor的失败状态，然后确定该怎么办, “怎么办”这回事叫做“supervisorStrategy"。
// Restart the storage child when StorageException is thrown.
// After 3 restarts within 5 seconds it will be stopped.
override val supervisorStrategy = OneForOneStrategy(maxNrOfRetries = 3,
withinTimeRange = 5 seconds) {
case _: Storage.StorageException => Restart
}
// After 3 restarts within 5 seconds it will be stopped.
override val supervisorStrategy = OneForOneStrategy(maxNrOfRetries = 3,
withinTimeRange = 5 seconds) {
case _: Storage.StorageException => Restart
}
SupervisorStrategy是干啥的呢？ An Akka Supervisor Strategy is the policy to apply for crashing children.
有哪些SupervisorStrategy可以用呢？
1. AllForOneStrategy : 当一个child失败以后，把所有的fault handling "Directive"应用于所有children。
2. OneForOneStrategy : 当一个child失败以后，把fault handling "Directive"应用于失败的那个child。
”Directive"是干啥的？
首先，这个东西需要决定当child failure之后，应该怎么办? “应该怎么办”是Akka提供了几个实现，表现在Directive的子类：
1. Resume 继续让失败的那个actor处理消息
2. Restart 抛弃旧的actor，然后创建一个新的，然后恢复消息处理
3. Stop 停止出问题的子actor
4. Escalate 把失败提升到父Actor的父Actor，通过重新跑出失败的cause，使用父Actor由于子actor同样的异常失败。
由此看，所谓的failure，就是异常，准确地说就是Throwable。
应该如何把Directive传给strategy?
OneForOneStrategy和 AllForOneStrategy是一个柯里化函数，第二个参数块是一个叫Decider的东西，它是一个PartialFunction，把异常和Directive对应起来。
type Decider = PartialFunction[Throwable, Directive]
在XXXStrategy的参数中，第一个参数块里的参数都可以省略，使得可以不限制重试次数，或者只限制总次数不限制频繁程度。
当没有指定supervisor strategy时，默认的行为是怎么样的呢？
• ActorInitializationException will stop the failing child actor
• ActorKilledException will stop the failing child actor
• Exception will restart the failing child actor
• Other types of Throwable will be escalated to parent actor
• ActorKilledException will stop the failing child actor
• Exception will restart the failing child actor
• Other types of Throwable will be escalated to parent actor
可以订制Strategy的行为：例子
import akka.actor.OneForOneStrategy
import akka.actor.SupervisorStrategy._
import scala.concurrent.duration._
override val supervisorStrategy =
OneForOneStrategy(maxNrOfRetries = 10, withinTimeRange = 1 minute) {
case _: ArithmeticException => Resume
case t =>
super.supervisorStrategy.decider.applyOrElse(t, (_: Any) => Escalate)
}
import akka.actor.SupervisorStrategy._
import scala.concurrent.duration._
override val supervisorStrategy =
OneForOneStrategy(maxNrOfRetries = 10, withinTimeRange = 1 minute) {
case _: ArithmeticException => Resume
case t =>
super.supervisorStrategy.decider.applyOrElse(t, (_: Any) => Escalate)
}
Supervision of Top-Level Actors
所有Top-Level Actors都是User Guardian的子actor。所以它们的失败由guardian根据它配置的strategy来处理
