iemejia commented on code in PR #12400: URL: https://github.com/apache/gluten/pull/12400#discussion_r3549010628
########## docs/get-started/VeloxLocalCache.md: ########## @@ -13,7 +13,7 @@ spark.gluten.sql.columnar.backend.velox.memCacheSize // the total size of i spark.gluten.sql.columnar.backend.velox.ssdCachePath // the folder to store the cache files, default is "/tmp". spark.gluten.sql.columnar.backend.velox.ssdCacheSize // the total size of the SSD cache, default is 128MB. Velox will do in-mem cache only if this value is 0. spark.gluten.sql.columnar.backend.velox.ssdCacheShards // the shards of the SSD cache, default is 1. -spark.gluten.sql.columnar.backend.velox.ssdCacheIOThreads // the IO threads for cache promoting, default is 1. Velox will try to do "read-ahead" if this value is bigger than 1 +spark.gluten.sql.columnar.backend.velox.ssdCacheIOThreads // the number of IO threads for SSD cache read/write operations, default is 4. Velox will try to do "read-ahead" if this value is bigger than 1 Review Comment: Fixed. Removed the trailing whitespace. ########## backends-velox/src/test/scala/org/apache/spark/sql/execution/VeloxFileHandleCacheSuite.scala: ########## @@ -0,0 +1,310 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.spark.sql.execution + +import org.apache.gluten.config.VeloxConfig +import org.apache.gluten.execution.{BasicScanExecTransformer, VeloxWholeStageTransformerSuite} + +import org.apache.spark.SparkConf + +/** + * Test suite for Velox file handle cache behavior. + * + * Tests correctness, config propagation, and edge cases for the file handle cache which caches open + * file handles (descriptors) to avoid repeated open/close overhead. + */ +class VeloxFileHandleCacheSuite extends VeloxWholeStageTransformerSuite { + override protected val resourcePath: String = "/parquet-for-read" + override protected val fileFormat: String = "parquet" + + override protected def sparkConf: SparkConf = { + super.sparkConf + .set(VeloxConfig.COLUMNAR_VELOX_FILE_HANDLE_CACHE_ENABLED.key, "true") + .set(VeloxConfig.COLUMNAR_VELOX_FILE_HANDLE_EXPIRATION_DURATION_MS.key, "2000") + .set(VeloxConfig.COLUMNAR_VELOX_NUM_CACHE_FILE_HANDLES.key, "10000") + } + + testWithSpecifiedSparkVersion( + "basic scan correctness with file handle cache enabled", + "3.5", + "3.5") { + // Verify that enabling file handle cache produces correct scan results + withTempPath { + dir => + spark + .range(10000) + .selectExpr("id", "cast(id % 7 as int) as category", "id * 1.5 as value") + .repartition(10) + .write + .parquet(dir.getCanonicalPath) + + val df = spark.read.parquet(dir.getCanonicalPath) + df.createOrReplaceTempView("t") + + runQueryAndCompare("SELECT count(*) FROM t") { + checkGlutenPlan[BasicScanExecTransformer] + } + runQueryAndCompare("SELECT sum(value) FROM t WHERE category = 3") { + checkGlutenPlan[BasicScanExecTransformer] + } + runQueryAndCompare("SELECT category, count(*) FROM t GROUP BY category") { + checkGlutenPlan[BasicScanExecTransformer] + } + } + } + + testWithSpecifiedSparkVersion( + "repeated scans produce consistent results (cache hit path)", + "3.5", + "3.5") { + // When file handles are cached, repeated scans of the same files must produce + // identical results. This exercises the cache hit path. Review Comment: Fixed. Reworded to focus on the actual invariant (result consistency across repeated scans) without claiming cache hits. ########## backends-velox/src/test/scala/org/apache/spark/sql/execution/VeloxFileHandleCacheSuite.scala: ########## @@ -0,0 +1,310 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.spark.sql.execution + +import org.apache.gluten.config.VeloxConfig +import org.apache.gluten.execution.{BasicScanExecTransformer, VeloxWholeStageTransformerSuite} + +import org.apache.spark.SparkConf + +/** + * Test suite for Velox file handle cache behavior. + * + * Tests correctness, config propagation, and edge cases for the file handle cache which caches open + * file handles (descriptors) to avoid repeated open/close overhead. + */ +class VeloxFileHandleCacheSuite extends VeloxWholeStageTransformerSuite { + override protected val resourcePath: String = "/parquet-for-read" + override protected val fileFormat: String = "parquet" + + override protected def sparkConf: SparkConf = { + super.sparkConf + .set(VeloxConfig.COLUMNAR_VELOX_FILE_HANDLE_CACHE_ENABLED.key, "true") + .set(VeloxConfig.COLUMNAR_VELOX_FILE_HANDLE_EXPIRATION_DURATION_MS.key, "2000") + .set(VeloxConfig.COLUMNAR_VELOX_NUM_CACHE_FILE_HANDLES.key, "10000") + } + + testWithSpecifiedSparkVersion( + "basic scan correctness with file handle cache enabled", + "3.5", + "3.5") { + // Verify that enabling file handle cache produces correct scan results + withTempPath { + dir => + spark + .range(10000) + .selectExpr("id", "cast(id % 7 as int) as category", "id * 1.5 as value") + .repartition(10) + .write + .parquet(dir.getCanonicalPath) + + val df = spark.read.parquet(dir.getCanonicalPath) + df.createOrReplaceTempView("t") + + runQueryAndCompare("SELECT count(*) FROM t") { + checkGlutenPlan[BasicScanExecTransformer] + } + runQueryAndCompare("SELECT sum(value) FROM t WHERE category = 3") { + checkGlutenPlan[BasicScanExecTransformer] + } + runQueryAndCompare("SELECT category, count(*) FROM t GROUP BY category") { + checkGlutenPlan[BasicScanExecTransformer] + } + } + } + + testWithSpecifiedSparkVersion( + "repeated scans produce consistent results (cache hit path)", + "3.5", + "3.5") { + // When file handles are cached, repeated scans of the same files must produce + // identical results. This exercises the cache hit path. + withTempPath { + dir => + spark + .range(5000) + .selectExpr("id", "cast(id as string) as name") + .repartition(50) // 50 files to exercise many cache entries + .write + .parquet(dir.getCanonicalPath) + + val path = dir.getCanonicalPath + val expected = spark.read.parquet(path).count() + assert(expected == 5000) + + // Verify scans go through Gluten/Velox + checkGlutenPlan[BasicScanExecTransformer](spark.read.parquet(path)) + + // Scan the same files multiple times - each should hit the cache Review Comment: Fixed. Reworded to reflect the actual assertion (consistent results) rather than assuming cache hits. -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
