Github user HeartSaVioR commented on a diff in the pull request:

    https://github.com/apache/spark/pull/21428#discussion_r191629272
  
    --- Diff: 
sql/core/src/test/scala/org/apache/spark/sql/streaming/continuous/shuffle/ContinuousShuffleSuite.scala
 ---
    @@ -288,4 +267,153 @@ class ContinuousShuffleReadSuite extends StreamTest {
         val thirdEpoch = rdd.compute(rdd.partitions(0), 
ctx).map(_.getUTF8String(0).toString).toSet
         assert(thirdEpoch == Set("writer1-row1", "writer2-row0"))
       }
    +
    +  test("one epoch") {
    +    val reader = new ContinuousShuffleReadRDD(sparkContext, numPartitions 
= 1)
    +    val writer = new RPCContinuousShuffleWriter(
    +      0, new HashPartitioner(1), Array(readRDDEndpoint(reader)))
    +
    +    writer.write(Iterator(1, 2, 3))
    +
    +    assert(readEpoch(reader) == Seq(1, 2, 3))
    +  }
    +
    +  test("multiple epochs") {
    +    val reader = new ContinuousShuffleReadRDD(sparkContext, numPartitions 
= 1)
    +    val writer = new RPCContinuousShuffleWriter(
    +      0, new HashPartitioner(1), Array(readRDDEndpoint(reader)))
    +
    +    writer.write(Iterator(1, 2, 3))
    +    writer.write(Iterator(4, 5, 6))
    +
    +    assert(readEpoch(reader) == Seq(1, 2, 3))
    +    assert(readEpoch(reader) == Seq(4, 5, 6))
    +  }
    +
    +  test("empty epochs") {
    +    val reader = new ContinuousShuffleReadRDD(sparkContext, numPartitions 
= 1)
    +    val writer = new RPCContinuousShuffleWriter(
    +      0, new HashPartitioner(1), Array(readRDDEndpoint(reader)))
    +
    +    writer.write(Iterator())
    +    writer.write(Iterator(1, 2))
    +    writer.write(Iterator())
    +    writer.write(Iterator())
    +    writer.write(Iterator(3, 4))
    +    writer.write(Iterator())
    +
    +    assert(readEpoch(reader) == Seq())
    +    assert(readEpoch(reader) == Seq(1, 2))
    +    assert(readEpoch(reader) == Seq())
    +    assert(readEpoch(reader) == Seq())
    +    assert(readEpoch(reader) == Seq(3, 4))
    +    assert(readEpoch(reader) == Seq())
    +  }
    +
    +  test("blocks waiting for writer") {
    +    val reader = new ContinuousShuffleReadRDD(sparkContext, numPartitions 
= 1)
    +    val writer = new RPCContinuousShuffleWriter(
    +      0, new HashPartitioner(1), Array(readRDDEndpoint(reader)))
    +
    +    val readerEpoch = reader.compute(reader.partitions(0), ctx)
    +
    +    val readRowThread = new Thread {
    +      override def run(): Unit = {
    +        assert(readerEpoch.toSeq.map(_.getInt(0)) == Seq(1))
    +      }
    +    }
    +    readRowThread.start()
    +
    +    eventually(timeout(streamingTimeout)) {
    +      assert(readRowThread.getState == Thread.State.TIMED_WAITING)
    +    }
    +
    +    // Once we write the epoch the thread should stop waiting and succeed.
    +    writer.write(Iterator(1))
    +    readRowThread.join()
    +  }
    +
    +  test("multiple writer partitions") {
    --- End diff --
    
    Would we want to have another test which covers out-of-order epoch between 
writers (if that's valid case for us), or rely on the test in 
ContinuousShuffleReadRDD?


---

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to