Digging into akka distributed workers scala

(rahat) #1

I was looking into WorkState.scala for the below example:

If I keep running the demo indefinitely, the acceptedWorkIds and doneWorkIds will keep growing as they are Set[String].
Besides recovery will take longer as no snapshot enabled. Now, say If I don’t need records older than 24 hours and set TTL for 24 hours for persistent journal, then problem would be, if for a particular completed Work, WorkAccepted(work) event records expires but rest stays, in that case if I just add
if(isAccepted(workId)) check for all update events, will that will be sufficient ?

    case WorkStarted(workId) => 
      if(isAccepted(workId)) {
        val (work, rest) = pendingWork.dequeue
        require(workId == work.workId, s"WorkStarted expected workId $workId == ${work.workId}")
          pendingWork = rest,
          workInProgress = workInProgress + (workId -> work))
      } else this