turboFei commented on a change in pull request #26339: [SPARK-27194][SPARK-29302][SQL] Fix the issue that for dynamic partition overwrite a task would conflict with its speculative task URL: https://github.com/apache/spark/pull/26339#discussion_r410834586
########## File path: sql/core/src/test/scala/org/apache/spark/sql/sources/PartitionedWriteSuite.scala ########## @@ -157,3 +161,49 @@ class PartitionedWriteSuite extends QueryTest with SharedSparkSession { } } } + +private class DetectDynamicSpeculationCommitProtocol( + jobId: String, + path: String, + dynamicPartitionOverwrite: Boolean) + extends HadoopMapReduceCommitProtocol(jobId, path, dynamicPartitionOverwrite) { + + override def commitTask(taskContext: TaskAttemptContext): TaskCommitMessage = { + if (dynamicPartitionOverwrite) { + val partitionPathSet = dynamicStagingTaskFiles + .map(taskFile => getDynamicPartitionPath(taskFile, taskContext)) + .map(_.toUri.getPath.stripPrefix(stagingDir.toUri.getPath + + Path.SEPARATOR)) + assert(partitionPathSet.equals(partitionPaths)) + } + super.commitTask(taskContext) + } +} + +class PartitionedSpeculateWriteSuite extends QueryTest with SharedSparkSession { + import testImplicits._ + + override def sparkConf(): SparkConf = { + super.sparkConf + .set(config.SPECULATION_MULTIPLIER, 0.0) + .set(config.SPECULATION_QUANTILE, 0.5) + .set(config.SPECULATION_ENABLED, true) + } + + test("SPARK-27194 SPARK-29302: Fix the issue that for dynamic partition overwrite, a " + + "task would conflict with its speculative task") { + withSQLConf(SQLConf.PARTITION_OVERWRITE_MODE.key -> PartitionOverwriteMode.DYNAMIC.toString, + SQLConf.FILE_COMMIT_PROTOCOL_CLASS.key -> + classOf[DetectDynamicSpeculationCommitProtocol].getName) { + withTable("t") { Review comment: It does pass without this fix. Because the local file system allows two task write the same file. But for distributed file system, it does not allow two task write the same file. An exception like `no lease on inode` thrown for DFS. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org