Finished a POC of Spark structured streaming from Kinesis as a source, it was working fine with Spark 2.4.5, that I had installed locally at the time.
Then at some point I had to upgrade to Spark 3.0.0 (also tested with 3.0.1) and this is when it started failing with this error:
Continuous processing does not support StreamingRelation operations.;;
kinesis
org.apache.spark.sql.AnalysisException: Continuous processing does not support StreamingRelation operations.;;
kinesis
at org.apache.spark.sql.catalyst.analysis.UnsupportedOperationChecker$.throwError(UnsupportedOperationChecker.scala:431)
at org.apache.spark.sql.catalyst.analysis.UnsupportedOperationChecker$.$anonfun$checkForContinuous$1(UnsupportedOperationChecker.scala:408)
at org.apache.spark.sql.catalyst.analysis.UnsupportedOperationChecker$.$anonfun$checkForContinuous$1$adapted(UnsupportedOperationChecker.scala:390)
at org.apache.spark.sql.catalyst.trees.TreeNode.foreachUp(TreeNode.scala:177)
at org.apache.spark.sql.catalyst.analysis.UnsupportedOperationChecker$.checkForContinuous(UnsupportedOperationChecker.scala:390)
at org.apache.spark.sql.streaming.StreamingQueryManager.createQuery(StreamingQueryManager.scala:290)
at org.apache.spark.sql.streaming.StreamingQueryManager.startQuery(StreamingQueryManager.scala:359)
at org.apache.spark.sql.streaming.DataStreamWriter.start(DataStreamWriter.scala:366)
at com.niceic.dl.kinesispuller.RunnerErr$.main(RunnerErrorSpark3.scala:50)
at com.niceic.dl.kinesispuller.Spark3ErrorTest.$anonfun$new$2(test.scala:32)
at org.scalatest.OutcomeOf.outcomeOf(OutcomeOf.scala:85)
at org.scalatest.OutcomeOf.outcomeOf$(OutcomeOf.scala:83)
at org.scalatest.OutcomeOf$.outcomeOf(OutcomeOf.scala:104)
at org.scalatest.Transformer.apply(Transformer.scala:22)
at org.scalatest.Transformer.apply(Transformer.scala:20)
at org.scalatest.funsuite.AnyFunSuiteLike$$anon$1.apply(AnyFunSuiteLike.scala:190)
at org.scalatest.TestSuite.withFixture(TestSuite.scala:196)
at org.scalatest.TestSuite.withFixture$(TestSuite.scala:195)
at org.scalatest.funsuite.AnyFunSuite.withFixture(AnyFunSuite.scala:1563)
at org.scalatest.funsuite.AnyFunSuiteLike.invokeWithFixture$1(AnyFunSuiteLike.scala:188)
at org.scalatest.funsuite.AnyFunSuiteLike.$anonfun$runTest$1(AnyFunSuiteLike.scala:200)
at org.scalatest.SuperEngine.runTestImpl(Engine.scala:306)
at org.scalatest.funsuite.AnyFunSuiteLike.runTest(AnyFunSuiteLike.scala:200)
at org.scalatest.funsuite.AnyFunSuiteLike.runTest$(AnyFunSuiteLike.scala:182)
at org.scalatest.funsuite.AnyFunSuite.runTest(AnyFunSuite.scala:1563)
at org.scalatest.funsuite.AnyFunSuiteLike.$anonfun$runTests$1(AnyFunSuiteLike.scala:233)
at org.scalatest.SuperEngine.$anonfun$runTestsInBranch$1(Engine.scala:413)
at scala.collection.immutable.List.foreach(List.scala:431)
at org.scalatest.SuperEngine.traverseSubNodes$1(Engine.scala:401)
at org.scalatest.SuperEngine.runTestsInBranch(Engine.scala:396)
at org.scalatest.SuperEngine.runTestsImpl(Engine.scala:475)
at org.scalatest.funsuite.AnyFunSuiteLike.runTests(AnyFunSuiteLike.scala:233)
at org.scalatest.funsuite.AnyFunSuiteLike.runTests$(AnyFunSuiteLike.scala:232)
at org.scalatest.funsuite.AnyFunSuite.runTests(AnyFunSuite.scala:1563)
at org.scalatest.Suite.run(Suite.scala:1112)
at org.scalatest.Suite.run$(Suite.scala:1094)
at org.scalatest.funsuite.AnyFunSuite.org$scalatest$funsuite$AnyFunSuiteLike$$super$run(AnyFunSuite.scala:1563)
at org.scalatest.funsuite.AnyFunSuiteLike.$anonfun$run$1(AnyFunSuiteLike.scala:237)
at org.scalatest.SuperEngine.runImpl(Engine.scala:535)
at org.scalatest.funsuite.AnyFunSuiteLike.run(AnyFunSuiteLike.scala:237)
at org.scalatest.funsuite.AnyFunSuiteLike.run$(AnyFunSuiteLike.scala:236)
at com.niceic.dl.kinesispuller.Spark3ErrorTest.org$scalatest$BeforeAndAfterAllConfigMap$$super$run(test.scala:24)
at org.scalatest.BeforeAndAfterAllConfigMap.liftedTree1$1(BeforeAndAfterAllConfigMap.scala:248)
at org.scalatest.BeforeAndAfterAllConfigMap.run(BeforeAndAfterAllConfigMap.scala:245)
at org.scalatest.BeforeAndAfterAllConfigMap.run$(BeforeAndAfterAllConfigMap.scala:242)
at com.niceic.dl.kinesispuller.Spark3ErrorTest.run(test.scala:24)
at org.scalatest.tools.SuiteRunner.run(SuiteRunner.scala:45)
at org.scalatest.tools.Runner$.$anonfun$doRunRunRunDaDoRunRun$13(Runner.scala:1320)
at org.scalatest.tools.Runner$.$anonfun$doRunRunRunDaDoRunRun$13$adapted(Runner.scala:1314)
at scala.collection.immutable.List.foreach(List.scala:431)
at org.scalatest.tools.Runner$.doRunRunRunDaDoRunRun(Runner.scala:1314)
at org.scalatest.tools.Runner$.$anonfun$runOptionallyWithPassFailReporter$24(Runner.scala:993)
at org.scalatest.tools.Runner$.$anonfun$runOptionallyWithPassFailReporter$24$adapted(Runner.scala:971)
at org.scalatest.tools.Runner$.withClassLoaderAndDispatchReporter(Runner.scala:1480)
at org.scalatest.tools.Runner$.runOptionallyWithPassFailReporter(Runner.scala:971)
at org.scalatest.tools.Runner$.run(Runner.scala:798)
at org.scalatest.tools.Runner.run(Runner.scala)
at org.jetbrains.plugins.scala.testingSupport.scalaTest.ScalaTestRunner.runScalaTest2or3(ScalaTestRunner.java:41)
at org.jetbrains.plugins.scala.testingSupport.scalaTest.ScalaTestRunner.main(ScalaTestRunner.java:28)
I run my code in IntelliJ (Spark installed locally with homebrew):
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.streaming.Trigger
object RunnerErr {
def main(): Unit = {
val spark = SparkSession
.builder
.appName("KinesisSpark3Error")
.master("local[*]")
.getOrCreate()
val kinesisSqlOptions = Map("streamName" -> "shards2"
, "endpointUrl" -> "https://kinesis.us-west-2.amazonaws.com"
, "awsAccessKeyId" -> "<awsAccessKeyId>"
, "awsSecretKey" -> "<awsSecretKey>"
, "startingPosition" -> "latest"
)
val df = spark.readStream
.format("kinesis")
.options(kinesisSqlOptions)
.load
df.writeStream
.outputMode("append")
.format("console")
.trigger(Trigger.Continuous("2 seconds")) //<-works on 2.4.5 only
// .trigger(Trigger.ProcessingTime("2 seconds")) //<-works on both
.start()
.awaitTermination()
}
}
I don't see any changes related to continuous trigger in Spark documentation between versions 2.4 and 3.0. The error is coming from UnsupportedOperationChecker.scala:
def checkForContinuous(plan: LogicalPlan, outputMode: OutputMode): Unit = {
checkForStreaming(plan, outputMode)
plan.foreachUp { implicit subPlan =>
subPlan match {
case (_: Project | _: Filter | _: MapElements | _: MapPartitions |
_: DeserializeToObject | _: SerializeFromObject | _: SubqueryAlias |
_: TypedFilter) =>
case node if node.nodeName == "StreamingRelationV2" =>
case Repartition(1, false, _) =>
case node: Aggregate =>
val aboveSinglePartitionCoalesce = node.find {
case Repartition(1, false, _) => true
case _ => false
}.isDefined
if (!aboveSinglePartitionCoalesce) {
throwError(s"In continuous processing mode, coalesce(1) must be called before " +
s"aggregate operation ${node.nodeName}.")
}
case node =>
throwError(s"Continuous processing does not support ${node.nodeName} operations.")
}
On the last case node of the snippet debugger shows node.sourceName = "kinesis". Again, works on Spark 2.4.5, fails on 3.0.0 and 3.0.1.
Kinesis Spark connector used:
- for Spark 2.4.5 is "com.qubole.spark" %% "spark-sql-kinesis" % "1.2.0_spark-2.4"
- for Spark 3+ is "com.qubole.spark" %% "spark-sql-kinesis" % "1.2.0_spark-3.0"
At this point I believe it's a bug - what do you think? Any input would be greatly appreciated - thanks!
Update1: Found in Spark source code that there was a change in DataStreamReader load method where for the kinesis source it returns StreamingRelation in Spark 3 and StreamingRelationV2 in Spark 2.4.5. Method checkForContinuous called on the dataframe plan later can handle StreamingRelationV2 and throws on StreamingRelation.
Update2: Qubole answered: "Spark 3 has significant changes in Data source V2 APIs which is required for continuous streaming. We had to remove the code for continuous streaming to support the connector in Spark 3. #92"
