Question

我尝试使用akka-http作为SSE来传输PersistenceQuery结果，但似乎当从客户端关闭http连接时，PersistenceQuery流仍然定期访问事件后端。

// Http part
complete {
   source(id)
        .map(e => e.event) // other transformations
        .map(e => ServerSentEvent(m.toString))
        .keepAlive(4 seconds, () => ServerSentEvent.heartbeat)
}

// source
def source(id: UUID)(implicit system: ActorSystem, materializer: ActorMaterializer)= {
    import system.dispatcher

    val journalQuery = PersistenceQuery(system).readJournalFor[CassandraReadJournal](CassandraReadJournal.Identifier)

    val futureSrcGraph:  RunnableGraph[Future[Source[EventEnvelope, NotUsed]]] =
      journalQuery.currentEventsByPersistenceId(id.toString, 0, Long.MaxValue)
        .map(_.sequenceNr)
        .toMat(Sink.last)(Keep.right)
        .mapMaterializedValue(fs => fs.recoverWith {
          case _ => Future { 0L } // assume we start at 1
        }.map(s => journalQuery.eventsByPersistenceId(id.toString, s + 1, Long.MaxValue)))

    Source.fromFutureSource(futureSrcGraph.run())

所以这基本上工作唯一的问题是流从未完成或看起来如此。我用CassandraReadJournal和LevelDb

尝试过

日志输出示例：

[DEBUG] [06/18/2018 10:52:16.774] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query from seqNr [6] in partition [0]
[DEBUG] [06/18/2018 10:52:16.790] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query took [15] ms (empty)
[DEBUG] [06/18/2018 10:52:16.790] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query from seqNr [6] in partition [1]
[DEBUG] [06/18/2018 10:52:16.796] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query took [5] ms (empty)
[DEBUG] [06/18/2018 10:52:19.768] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query from seqNr [6] in partition [0]
[DEBUG] [06/18/2018 10:52:19.784] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query took [15] ms (empty)
[DEBUG] [06/18/2018 10:52:19.784] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query from seqNr [6] in partition [1]
[DEBUG] [06/18/2018 10:52:19.790] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query took [6] ms (empty)
[DEBUG] [06/18/2018 10:52:22.765] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query from seqNr [6] in partition [0]
[DEBUG] [06/18/2018 10:52:22.772] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query took [6] ms (empty)
[DEBUG] [06/18/2018 10:52:22.772] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query from seqNr [6] in partition [1]
[DEBUG] [06/18/2018 10:52:22.790] [sys-cassandra-plugin-default-dispatcher-17] [EventsByPersistenceIdStage(akka://sys)] EventsByPersistenceId [c6031a8a-db71-4dcb-9d4f-f140faa2f4c4] Query took [17] ms (empty)

它永远持续下去。

我也试过省略Source.fromFutureSource，只是运行带有相同结果的journalQuery.eventsByPersistenceId。

我做错了什么？

Answer 1

以下是即使客户端关闭连接，我的公司代理也不会断开与服务器的连接。

PersistenceQuery和Akka-Http僵尸流

1 个答案: