Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/18092#discussion_r121269114
--- Diff:
core/src/test/scala/org/apache/spark/storage/BlockManagerSuite.scala ---
@@ -1281,6 +1285,57 @@ class BlockManagerSuite extends SparkFunSuite with
Matchers with BeforeAndAfterE
assert(master.getLocations("item").isEmpty)
}
+ test("SPARK-20640: Shuffle registration timeout and maxAttempts conf are
working") {
+ val shufflePort = 10000
+ val tryAgainMsg = "test_spark_20640_try_again"
+ conf.set("spark.shuffle.service.enabled", "true")
+ conf.set("spark.shuffle.service.port", shufflePort.toString)
+ // a server which delays response 50ms and must try twice for success.
+ def newShuffleServer(): TransportServer = {
+ val attempts = new mutable.HashMap[String, Int]()
+ val handler = new NoOpRpcHandler {
+ override def receive(client: TransportClient, message: ByteBuffer,
+ callback: RpcResponseCallback): Unit = {
+ val msgObj = BlockTransferMessage.Decoder.fromByteBuffer(message)
+ msgObj match {
+ case exec: RegisterExecutor =>
+ Thread.sleep(50)
--- End diff --
Do we actually need this sleep? What if we just simply never returned any
response if `attempt < 2`?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
