Improve exception handling when adding artifacts

Author: Hendrik Huebner

sql/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectAddArtifactsHandler.scala

@@ -29,6 +29,7 @@ import io.grpc.stub.StreamObserver
 import org.apache.spark.connect.proto
 import org.apache.spark.connect.proto.{AddArtifactsRequest, AddArtifactsResponse}
 import org.apache.spark.connect.proto.AddArtifactsResponse.ArtifactSummary
+import org.apache.spark.SparkRuntimeException
 import org.apache.spark.sql.artifact.ArtifactManager
 import org.apache.spark.sql.connect.utils.ErrorUtils
 import org.apache.spark.sql.util.ArtifactUtils
@@ -112,19 +113,34 @@ class SparkConnectAddArtifactsHandler(val responseObserver: StreamObserver[AddAr
    * @return
    */
   protected def flushStagedArtifacts(): Seq[ArtifactSummary] = {
+    val failedArtifactExceptions = mutable.ListBuffer[SparkRuntimeException]()
+
     // Non-lazy transformation when using Buffer.
-    stagedArtifacts.map { artifact =>
-      // We do not store artifacts that fail the CRC. The failure is reported in the artifact
-      // summary and it is up to the client to decide whether to retry sending the artifact.
-      if (artifact.getCrcStatus.contains(true)) {
-        if (artifact.path.startsWith(ArtifactManager.forwardToFSPrefix + File.separator)) {
-          holder.artifactManager.uploadArtifactToFs(artifact.path, artifact.stagedPath)
-        } else {
-          addStagedArtifactToArtifactManager(artifact)
+    val summaries = stagedArtifacts.map { artifact =>
+      try {
+        // We do not store artifacts that fail the CRC. The failure is reported in the artifact
+        // summary and it is up to the client to decide whether to retry sending the artifact.
+        if (artifact.getCrcStatus.contains(true)) {
+          if (artifact.path.startsWith(ArtifactManager.forwardToFSPrefix + File.separator)) {
+            holder.artifactManager.uploadArtifactToFs(artifact.path, artifact.stagedPath)
+          } else {
+            addStagedArtifactToArtifactManager(artifact)
+          }
         }
+      } catch {
+        case e: SparkRuntimeException if e.getCondition == "ARTIFACT_ALREADY_EXISTS" =>
+          failedArtifactExceptions += e
       }
       artifact.summary()
     }.toSeq
+
+    if (failedArtifactExceptions.nonEmpty) {
+      val exception = failedArtifactExceptions.head
+      failedArtifactExceptions.drop(1).foreach(exception.addSuppressed(_))
+      throw exception
+    }
+
+    summaries
   }
 
   protected def cleanUpStagedArtifacts(): Unit = Utils.deleteRecursively(stagingDir.toFile)

sql/connect/server/src/test/scala/org/apache/spark/sql/connect/service/AddArtifactsHandlerSuite.scala

@@ -34,6 +34,7 @@ import io.grpc.stub.StreamObserver
 
 import org.apache.spark.connect.proto
 import org.apache.spark.connect.proto.{AddArtifactsRequest, AddArtifactsResponse}
+import org.apache.spark.SparkRuntimeException
 import org.apache.spark.sql.connect.ResourceHelper
 import org.apache.spark.sql.test.SharedSparkSession
 import org.apache.spark.util.ThreadUtils
@@ -51,17 +52,31 @@ class AddArtifactsHandlerSuite extends SharedSparkSession with ResourceHelper {
     override def onCompleted(): Unit = {}
   }
 
-  class TestAddArtifactsHandler(responseObserver: StreamObserver[AddArtifactsResponse])
+  class TestAddArtifactsHandler(responseObserver: StreamObserver[AddArtifactsResponse],
+                                throwIfArtifactExists: Boolean = false)
       extends SparkConnectAddArtifactsHandler(responseObserver) {
 
     // Stop the staged artifacts from being automatically deleted
     override protected def cleanUpStagedArtifacts(): Unit = {}
 
     private val finalArtifacts = mutable.Buffer.empty[String]
+    private val artifactChecksums: mutable.Map[String, Long] = mutable.Map.empty
 
     // Record the artifacts that are sent out for final processing.
     override protected def addStagedArtifactToArtifactManager(artifact: StagedArtifact): Unit = {
+      // Throw if artifact already exists and has different checksum
+      // This mocks the behavior of ArtifactManager.addArtifact without comparing the entire file
+      if (throwIfArtifactExists
+          && finalArtifacts.contains(artifact.name)
+          && artifact.getCrc != artifactChecksums(artifact.name)) {
+        throw new SparkRuntimeException(
+          "ARTIFACT_ALREADY_EXISTS",
+          Map("normalizedRemoteRelativePath" -> artifact.name)
+        )
+      }
+
       finalArtifacts.append(artifact.name)
+      artifactChecksums += (artifact.name -> artifact.getCrc)
     }
 
     def getFinalArtifacts: Seq[String] = finalArtifacts.toSeq
@@ -418,4 +433,44 @@ class AddArtifactsHandlerSuite extends SharedSparkSession with ResourceHelper {
     }
   }
 
+  test("All artifacts are added, even if some fail") {
+    val promise = Promise[AddArtifactsResponse]()
+    val handler = new TestAddArtifactsHandler(new DummyAddArtifactsStreamObserver(promise),
+                                              throwIfArtifactExists = true)
+    try {
+      val name1 = "jars/dummy1.jar"
+      val name2 = "jars/dummy2.jar"
+      val name3 = "jars/dummy3.jar"
+
+      val artifactPath1 = inputFilePath.resolve("smallClassFile.class")
+      val artifactPath2 = inputFilePath.resolve("smallJar.jar")
+
+      assume(artifactPath1.toFile.exists)
+      addSingleChunkArtifact(handler, sessionKey, name1, artifactPath1)
+      addSingleChunkArtifact(handler, sessionKey, name3, artifactPath1)
+
+      val e = intercept[StatusRuntimeException] {
+        addSingleChunkArtifact(handler, sessionKey, name1, artifactPath2)
+        addSingleChunkArtifact(handler, sessionKey, name2, artifactPath1)
+        addSingleChunkArtifact(handler, sessionKey, name3, artifactPath2)
+        handler.onCompleted()
+      }
+
+      // Both artifacts should be added, despite exception
+      assert(handler.getFinalArtifacts.contains(name1))
+      assert(handler.getFinalArtifacts.contains(name2))
+      assert(handler.getFinalArtifacts.contains(name3))
+
+      assert(e.getStatus.getCode == Code.INTERNAL)
+      val statusProto = StatusProto.fromThrowable(e)
+      assert(statusProto.getDetailsCount == 1)
+      val details = statusProto.getDetails(0)
+      val info = details.unpack(classOf[ErrorInfo])
+
+      assert(e.getMessage.contains("ARTIFACT_ALREADY_EXISTS"))
+      assert(info.getMetadataMap().get("messageParameters").contains(name1))
+    } finally {
+      handler.forceCleanUp()
+    }
+  }
 }

sql/core/src/main/scala/org/apache/spark/sql/artifact/ArtifactManager.scala

@@ -25,6 +25,7 @@ import java.nio.file.{CopyOption, Files, Path, Paths, StandardCopyOption}
 import java.util.concurrent.CopyOnWriteArrayList
 import java.util.concurrent.atomic.AtomicBoolean
 
+import scala.collection.mutable.ListBuffer
 import scala.jdk.CollectionConverters._
 import scala.reflect.ClassTag
 
@@ -266,28 +267,41 @@ class ArtifactManager(session: SparkSession) extends AutoCloseable with Logging
    * they are from a permanent location.
    */
   private[sql] def addLocalArtifacts(artifacts: Seq[Artifact]): Unit = {
+    val failedArtifactExceptions = ListBuffer[RuntimeException]()
+
     artifacts.foreach { artifact =>
-      artifact.storage match {
-        case d: Artifact.LocalFile =>
-          addArtifact(
-            artifact.path,
-            d.path,
-            fragment = None,
-            deleteStagedFile = false)
-        case d: Artifact.InMemory =>
-          val tempDir = Utils.createTempDir().toPath
-          val tempFile = tempDir.resolve(artifact.path.getFileName)
-          val outStream = Files.newOutputStream(tempFile)
-          Utils.tryWithSafeFinallyAndFailureCallbacks {
-            d.stream.transferTo(outStream)
-            addArtifact(artifact.path, tempFile, fragment = None)
-          }(finallyBlock = {
-            outStream.close()
-          })
-        case _ =>
-          throw SparkException.internalError(s"Unsupported artifact storage: ${artifact.storage}")
+      try {
+        artifact.storage match {
+          case d: Artifact.LocalFile =>
+            addArtifact(
+              artifact.path,
+              d.path,
+              fragment = None,
+              deleteStagedFile = false)
+          case d: Artifact.InMemory =>
+            val tempDir = Utils.createTempDir().toPath
+            val tempFile = tempDir.resolve(artifact.path.getFileName)
+            val outStream = Files.newOutputStream(tempFile)
+            Utils.tryWithSafeFinallyAndFailureCallbacks {
+              d.stream.transferTo(outStream)
+              addArtifact(artifact.path, tempFile, fragment = None)
+            }(finallyBlock = {
+              outStream.close()
+            })
+          case _ =>
+            throw SparkException.internalError(s"Unsupported artifact storage: ${artifact.storage}")
+        }
+      } catch {
+        case e: SparkRuntimeException =>
+          failedArtifactExceptions += e
       }
     }
+
+    if (failedArtifactExceptions.nonEmpty) {
+      val exception = failedArtifactExceptions.head
+      failedArtifactExceptions.drop(1).foreach(exception.addSuppressed(_))
+      throw exception
+    }
   }
 
   def classloader: ClassLoader = synchronized {

sql/core/src/test/scala/org/apache/spark/sql/artifact/ArtifactManagerSuite.scala

@@ -20,8 +20,9 @@ import java.io.File
 import java.nio.charset.StandardCharsets
 import java.nio.file.{Files, Path, Paths}
 
-import org.apache.spark.{SparkConf, SparkException}
+import org.apache.spark.{SparkConf, SparkException, SparkRuntimeException}
 import org.apache.spark.metrics.source.CodegenMetrics
+import org.apache.spark.sql.Artifact
 import org.apache.spark.sql.classic.SparkSession
 import org.apache.spark.sql.functions.col
 import org.apache.spark.sql.internal.SQLConf
@@ -346,6 +347,71 @@ class ArtifactManagerSuite extends SharedSparkSession {
     }
   }
 
+  test("Add multiple artifacts to local session and check if all are added despite exception") {
+    val copyDir = Utils.createTempDir().toPath
+    Utils.copyDirectory(artifactPath.toFile, copyDir.toFile)
+
+    val artifact1Path = "my/custom/pkg/artifact1.jar"
+    val artifact2Path = "my/custom/pkg/artifact2.jar"
+    val targetPath = Paths.get(artifact1Path)
+    val targetPath2 = Paths.get(artifact2Path)
+
+    val classPath1 = copyDir.resolve("Hello.class")
+    val classPath2 = copyDir.resolve("smallJar.jar")
+    assume(artifactPath.resolve("Hello.class").toFile.exists)
+    assume(artifactPath.resolve("smallClassFile.class").toFile.exists)
+
+    val artifact1 = Artifact.newArtifactFromExtension(
+      targetPath.getFileName.toString,
+      targetPath,
+      new Artifact.LocalFile(Paths.get(classPath1.toString)))
+
+    val alreadyExistingArtifact = Artifact.newArtifactFromExtension(
+      targetPath2.getFileName.toString,
+      targetPath,
+      new Artifact.LocalFile(Paths.get(classPath2.toString)))
+
+    val artifact2 = Artifact.newArtifactFromExtension(
+      targetPath2.getFileName.toString,
+      targetPath2,
+      new Artifact.LocalFile(Paths.get(classPath2.toString)))
+
+    spark.artifactManager.addLocalArtifacts(Seq(artifact1))
+
+    val exception = intercept[SparkRuntimeException] {
+      spark.artifactManager.addLocalArtifacts(
+        Seq(alreadyExistingArtifact, artifact2, alreadyExistingArtifact))
+    }
+
+    // Validate exception: Should be ARTIFACT_ALREADY_EXISTS and have one suppressed exception
+    assert(exception.getCondition == "ARTIFACT_ALREADY_EXISTS",
+      s"Expected ARTIFACT_ALREADY_EXISTS but got: ${exception.getCondition}")
+
+    assert(exception.getSuppressed.length == 1)
+    assert(exception.getSuppressed.head.isInstanceOf[SparkRuntimeException])
+    val suppressed = exception.getSuppressed.head.asInstanceOf[SparkRuntimeException]
+    assert(suppressed.getCondition == "ARTIFACT_ALREADY_EXISTS")
+
+    // Artifact1 should have been added
+    val expectedFile1 = ArtifactManager.artifactRootDirectory
+      .resolve(s"$sessionUUID/jars/$artifact1Path")
+      .toFile
+    assert(expectedFile1.exists())
+
+    // Artifact2 should have been added despite exception
+    val expectedFile2 = ArtifactManager.artifactRootDirectory
+      .resolve(s"$sessionUUID/jars/$artifact2Path")
+      .toFile
+    assert(expectedFile2.exists())
+
+    // Cleanup
+    artifactManager.cleanUpResourcesForTesting()
+    val sessionDir = ArtifactManager.artifactRootDirectory.resolve(sessionUUID).toFile
+
+    assert(!expectedFile1.exists())
+    assert(!sessionDir.exists())
+  }
+
   test("Added artifact can be loaded by the current SparkSession") {
     val path = artifactPath.resolve("IntSumUdf.class")
     assume(path.toFile.exists)