[Bugfix] Insights on query execution error (opensearch-project#475)

* BugFix: Add error logs Signed-off-by: Louis Chu <[email protected]> * Add IT Signed-off-by: Louis Chu <[email protected]> * Fix IT Signed-off-by: Louis Chu <[email protected]> * Log stacktrace Signed-off-by: Louis Chu <[email protected]> * Use full msg instead of prefix Signed-off-by: Louis Chu <[email protected]> --------- Signed-off-by: Louis Chu <[email protected]>
dai-chen · Jul 26, 2024 · 3c8a490 · 3c8a490
1 parent 538dd54
commit 3c8a490
Show file tree

Hide file tree

Showing 2 changed files with 75 additions and 10 deletions.
diff --git a/integ-test/src/test/scala/org/apache/spark/sql/FlintREPLITSuite.scala b/integ-test/src/test/scala/org/apache/spark/sql/FlintREPLITSuite.scala
@@ -422,6 +422,69 @@ class FlintREPLITSuite extends SparkFunSuite with OpenSearchSuite with JobTest {
     }
   }
 
+  test("create table with dummy location should fail with excepted error message") {
+    try {
+      createSession(jobRunId, "")
+      threadLocalFuture.set(startREPL())
+
+      val dummyLocation = "s3://path/to/dummy/location"
+      val testQueryId = "110"
+      val createTableStatement =
+        s"""
+           | CREATE TABLE $testTable
+           | (
+           |   name STRING,
+           |   age INT
+           | )
+           | USING CSV
+           | LOCATION '$dummyLocation'
+           | OPTIONS (
+           |  header 'false',
+           |  delimiter '\\t'
+           | )
+           |""".stripMargin
+      val createTableStatementId =
+        submitQuery(s"${makeJsonCompliant(createTableStatement)}", testQueryId)
+
+      val createTableStatementValidation: REPLResult => Boolean = result => {
+        assert(
+          result.results.size == 0,
+          s"expected result size is 0, but got ${result.results.size}")
+        assert(
+          result.schemas.size == 0,
+          s"expected schema size is 0, but got ${result.schemas.size}")
+        failureValidation(result)
+        true
+      }
+      pollForResultAndAssert(createTableStatementValidation, testQueryId)
+      assert(
+        !awaitConditionForStatementOrTimeout(
+          statement => {
+            statement.error match {
+              case Some(error)
+                  if error == """{"Message":"Fail to run query. Cause: No FileSystem for scheme \"s3\""}""" =>
+              // Assertion passed
+              case _ =>
+                fail(s"Statement error is: ${statement.error}")
+            }
+            statement.state == "failed"
+          },
+          createTableStatementId),
+        s"Fail to verify for $createTableStatementId.")
+      // clean up
+      val dropStatement =
+        s"""DROP TABLE $testTable""".stripMargin
+      submitQuery(s"${makeJsonCompliant(dropStatement)}", "999")
+    } catch {
+      case e: Exception =>
+        logError("Unexpected exception", e)
+        assert(false, "Unexpected exception")
+    } finally {
+      waitREPLStop(threadLocalFuture.get())
+      threadLocalFuture.remove()
+    }
+  }
+
   /**
    * JSON does not support raw newlines (\n) in string values. All newlines must be escaped or
    * removed when inside a JSON string. The same goes for tab characters, which should be

diff --git a/spark-sql-application/src/main/scala/org/apache/spark/sql/FlintJobExecutor.scala b/spark-sql-application/src/main/scala/org/apache/spark/sql/FlintJobExecutor.scala
@@ -436,18 +436,22 @@ trait FlintJobExecutor {
 
   private def handleQueryException(
       e: Exception,
-      message: String,
+      messagePrefix: String,
       errorSource: Option[String] = None,
       statusCode: Option[Int] = None): String = {
-
-    val errorDetails = Map("Message" -> s"$message: ${e.getMessage}") ++
+    val errorMessage = s"$messagePrefix: ${e.getMessage}"
+    val errorDetails = Map("Message" -> errorMessage) ++
       errorSource.map("ErrorSource" -> _) ++
       statusCode.map(code => "StatusCode" -> code.toString)
 
     val errorJson = mapper.writeValueAsString(errorDetails)
 
-    statusCode.foreach { code =>
-      CustomLogging.logError(new OperationMessage("", code), e)
+    // CustomLogging will call log4j logger.error() underneath
+    statusCode match {
+      case Some(code) =>
+        CustomLogging.logError(new OperationMessage(errorMessage, code), e)
+      case None =>
+        CustomLogging.logError(errorMessage, e)
     }
 
     errorJson
@@ -491,16 +495,14 @@ trait FlintJobExecutor {
       case r: SparkException =>
         handleQueryException(r, ExceptionMessages.SparkExceptionErrorPrefix)
       case r: Exception =>
-        val rootCauseClassName = ex.getClass.getName
-        val errMsg = ex.getMessage
-        logDebug(s"Root cause class name: $rootCauseClassName")
-        logDebug(s"Root cause error message: $errMsg")
+        val rootCauseClassName = r.getClass.getName
+        val errMsg = r.getMessage
         if (rootCauseClassName == "org.apache.hadoop.hive.metastore.api.MetaException" &&
           errMsg.contains("com.amazonaws.services.glue.model.AccessDeniedException")) {
           val e = new SecurityException(ExceptionMessages.GlueAccessDeniedMessage)
           handleQueryException(e, ExceptionMessages.QueryRunErrorPrefix)
         } else {
-          handleQueryException(ex, ExceptionMessages.QueryRunErrorPrefix)
+          handleQueryException(r, ExceptionMessages.QueryRunErrorPrefix)
         }
     }
   }