steveloughran commented on code in PR #37474: URL: https://github.com/apache/spark/pull/37474#discussion_r943267264
########## hadoop-cloud/src/hadoop-3/main/scala/org/apache/spark/internal/io/cloud/AbortableStreamBasedCheckpointFileManager.scala: ########## @@ -0,0 +1,94 @@ +/* + * 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.internal.io.cloud + +import java.nio.file.FileAlreadyExistsException +import java.util.EnumSet + +import scala.util.control.NonFatal + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs._ + +import org.apache.spark.internal.Logging +import org.apache.spark.sql.execution.streaming.AbstractFileContextBasedCheckpointFileManager +import org.apache.spark.sql.execution.streaming.CheckpointFileManager.CancellableFSDataOutputStream + +class AbortableStreamBasedCheckpointFileManager(path: Path, hadoopConf: Configuration) + extends AbstractFileContextBasedCheckpointFileManager(path, hadoopConf) with Logging { + + if (!fc.hasPathCapability(path, CommonPathCapabilities.ABORTABLE_STREAM)) { + throw new UnsupportedFileSystemException("AbortableStreamBasedCheckpointFileManager requires" + + " an fs with abortable stream support") + } + + logInfo(s"Writing atomically to $path based on abortable stream") + + class AbortableStreamBasedFSDataOutputStream( + fsDataOutputStream: FSDataOutputStream, + fc: FileContext, + path: Path, + overwriteIfPossible: Boolean) extends CancellableFSDataOutputStream(fsDataOutputStream) { + + @volatile private var terminated = false + + override def cancel(): Unit = synchronized { + if (terminated) return + try { + fsDataOutputStream.abort() + fsDataOutputStream.close() + } catch { + case NonFatal(e) => + logWarning(s"Error cancelling write to $path", e) + } finally { + terminated = true + } + } + + override def close(): Unit = synchronized { + if (terminated) return + try { + if (!overwriteIfPossible && fc.util().exists(path)) { + throw new FileAlreadyExistsException( + s"Failed to close atomic stream $path as destination already exists") + } + fsDataOutputStream.close() + } catch { + case NonFatal(e) => + logWarning(s"Error closing $path", e) Review Comment: again. log output stream ########## sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/CheckpointFileManagerSuite.scala: ########## @@ -58,50 +69,40 @@ abstract class CheckpointFileManagerTests extends SparkFunSuite with SQLHelper { // Create atomic without overwrite var path = new Path(s"$dir/file") assert(!fm.exists(path)) - fm.createAtomic(path, overwriteIfPossible = false).cancel() + fm.createAtomic(path, overwriteIfPossible = false).writeContent(1).cancel() assert(!fm.exists(path)) - fm.createAtomic(path, overwriteIfPossible = false).close() + fm.createAtomic(path, overwriteIfPossible = false).writeContent(2).close() assert(fm.exists(path)) + assert(fm.open(path).readInt() == 2) quietly { intercept[IOException] { // should throw exception since file exists and overwrite is false - fm.createAtomic(path, overwriteIfPossible = false).close() + fm.createAtomic(path, overwriteIfPossible = false).writeContent(3).close() } } + assert(fm.open(path).readInt() == 2) // Create atomic with overwrite if possible path = new Path(s"$dir/file2") assert(!fm.exists(path)) - fm.createAtomic(path, overwriteIfPossible = true).cancel() + fm.createAtomic(path, overwriteIfPossible = true).writeContent(4).cancel() assert(!fm.exists(path)) - fm.createAtomic(path, overwriteIfPossible = true).close() + fm.createAtomic(path, overwriteIfPossible = true).writeContent(5).close() assert(fm.exists(path)) - fm.createAtomic(path, overwriteIfPossible = true).close() // should not throw exception - - // crc file should not be leaked when origin file doesn't exist. - // The implementation of Hadoop filesystem may filter out checksum file, so - // listing files from local filesystem. - val fileNames = new File(path.getParent.toString).listFiles().toSeq - .filter(p => p.isFile).map(p => p.getName) - val crcFiles = fileNames.filter(n => n.startsWith(".") && n.endsWith(".crc")) - val originFileNamesForExistingCrcFiles = crcFiles.map { name => - // remove first "." and last ".crc" - name.substring(1, name.length - 4) - } - - // Check all origin files exist for all crc files. - assert(originFileNamesForExistingCrcFiles.toSet.subsetOf(fileNames.toSet), - s"Some of origin files for crc files don't exist - crc files: $crcFiles / " + - s"expected origin files: $originFileNamesForExistingCrcFiles / actual files: $fileNames") + assert(fm.open(path).readInt() == 5) + // should not throw exception + fm.createAtomic(path, overwriteIfPossible = true).writeContent(6).close() + assert(fm.open(path).readInt() == 6) Review Comment: be good to close this; may use up process resources for rest of test run ########## hadoop-cloud/src/hadoop-3/main/scala/org/apache/spark/internal/io/cloud/AbortableStreamBasedCheckpointFileManager.scala: ########## @@ -0,0 +1,94 @@ +/* + * 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.internal.io.cloud + +import java.nio.file.FileAlreadyExistsException +import java.util.EnumSet + +import scala.util.control.NonFatal + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs._ + +import org.apache.spark.internal.Logging +import org.apache.spark.sql.execution.streaming.AbstractFileContextBasedCheckpointFileManager +import org.apache.spark.sql.execution.streaming.CheckpointFileManager.CancellableFSDataOutputStream + +class AbortableStreamBasedCheckpointFileManager(path: Path, hadoopConf: Configuration) + extends AbstractFileContextBasedCheckpointFileManager(path, hadoopConf) with Logging { + + if (!fc.hasPathCapability(path, CommonPathCapabilities.ABORTABLE_STREAM)) { + throw new UnsupportedFileSystemException("AbortableStreamBasedCheckpointFileManager requires" + + " an fs with abortable stream support") + } + + logInfo(s"Writing atomically to $path based on abortable stream") + + class AbortableStreamBasedFSDataOutputStream( + fsDataOutputStream: FSDataOutputStream, + fc: FileContext, + path: Path, + overwriteIfPossible: Boolean) extends CancellableFSDataOutputStream(fsDataOutputStream) { + + @volatile private var terminated = false + + override def cancel(): Unit = synchronized { + if (terminated) return + try { + fsDataOutputStream.abort() + fsDataOutputStream.close() + } catch { + case NonFatal(e) => + logWarning(s"Error cancelling write to $path", e) + } finally { + terminated = true + } + } + + override def close(): Unit = synchronized { + if (terminated) return + try { + if (!overwriteIfPossible && fc.util().exists(path)) { + throw new FileAlreadyExistsException( Review Comment: maybe call abort() here for maximum rigorousness ########## hadoop-cloud/src/hadoop-3/main/scala/org/apache/spark/internal/io/cloud/AbortableStreamBasedCheckpointFileManager.scala: ########## @@ -0,0 +1,94 @@ +/* + * 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.internal.io.cloud + +import java.nio.file.FileAlreadyExistsException +import java.util.EnumSet + +import scala.util.control.NonFatal + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs._ + +import org.apache.spark.internal.Logging +import org.apache.spark.sql.execution.streaming.AbstractFileContextBasedCheckpointFileManager +import org.apache.spark.sql.execution.streaming.CheckpointFileManager.CancellableFSDataOutputStream + +class AbortableStreamBasedCheckpointFileManager(path: Path, hadoopConf: Configuration) + extends AbstractFileContextBasedCheckpointFileManager(path, hadoopConf) with Logging { + + if (!fc.hasPathCapability(path, CommonPathCapabilities.ABORTABLE_STREAM)) { + throw new UnsupportedFileSystemException("AbortableStreamBasedCheckpointFileManager requires" + + " an fs with abortable stream support") + } + + logInfo(s"Writing atomically to $path based on abortable stream") + + class AbortableStreamBasedFSDataOutputStream( + fsDataOutputStream: FSDataOutputStream, + fc: FileContext, + path: Path, + overwriteIfPossible: Boolean) extends CancellableFSDataOutputStream(fsDataOutputStream) { + + @volatile private var terminated = false + + override def cancel(): Unit = synchronized { + if (terminated) return + try { + fsDataOutputStream.abort() + fsDataOutputStream.close() + } catch { + case NonFatal(e) => + logWarning(s"Error cancelling write to $path", e) Review Comment: include SfsDataOutputStream in the log to get any info from it -- 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]
