HeartSaVioR commented on a change in pull request #26416: [SPARK-29779][CORE] 
Compact old event log files and cleanup
URL: https://github.com/apache/spark/pull/26416#discussion_r357463341
 
 

 ##########
 File path: 
core/src/main/scala/org/apache/spark/deploy/history/EventLogFileCompactor.scala
 ##########
 @@ -0,0 +1,157 @@
+/*
+ * 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.deploy.history
+
+import java.io.IOException
+import java.net.URI
+
+import org.apache.hadoop.conf.Configuration
+import org.apache.hadoop.fs.{FileStatus, FileSystem, Path}
+
+import org.apache.spark.SparkConf
+import org.apache.spark.internal.Logging
+import org.apache.spark.internal.config.EVENT_LOG_ROLLING_MAX_FILES_TO_RETAIN
+import org.apache.spark.scheduler._
+
+/**
+ * This class compacts the old event log files into one compact file, via two 
phases reading:
+ *
+ * 1) Initialize available [[EventFilterBuilder]] instances, and replay the 
old event log files with
+ * builders, so that these builders can gather the information to create 
[[EventFilter]] instances.
+ * 2) Initialize [[EventFilter]] instances from [[EventFilterBuilder]] 
instances, and replay the
+ * old event log files with filters. Rewrite the events to the compact file 
which the filters decide
+ * to accept.
+ *
+ * This class assumes caller will provide the sorted list of files which are 
sorted by the index of
+ * event log file - caller should keep in mind that this class doesn't care 
about the semantic of
+ * ordering.
+ *
+ * When compacting the files, the range of compaction for given file list is 
determined as:
+ * (rightmost compact file ~ the file where there're `maxFilesToRetain` files 
on the right side)
+ *
+ * If there's no compact file in the list, it starts from the first file. If 
there're not enough
+ * files after rightmost compact file, compaction will be skipped.
+ */
+class EventLogFileCompactor(
+    sparkConf: SparkConf,
+    hadoopConf: Configuration,
+    fs: FileSystem) extends Logging {
+
+  private val maxFilesToRetain: Int = 
sparkConf.get(EVENT_LOG_ROLLING_MAX_FILES_TO_RETAIN)
+
+  def compact(eventLogFiles: Seq[FileStatus]): Seq[FileStatus] = {
+    if (eventLogFiles.length <= maxFilesToRetain) {
+      return eventLogFiles
+    }
+
+    if (EventLogFileWriter.isCompacted(eventLogFiles.last.getPath)) {
+      return Seq(eventLogFiles.last)
+    }
+
+    val (filesToCompact, filesToRetain) = findFilesToCompact(eventLogFiles)
+    if (filesToCompact.isEmpty) {
+      filesToRetain
+    } else {
+      val builders = EventFilterBuilder.initializeBuilders(fs, 
filesToCompact.map(_.getPath))
+
+      val rewriter = new FilteredEventLogFileRewriter(sparkConf, hadoopConf, 
fs,
+        builders.map(_.createFilter()))
+      val compactedPath = rewriter.rewrite(filesToCompact)
+
+      cleanupCompactedFiles(filesToCompact)
+
+      fs.getFileStatus(new Path(compactedPath)) :: filesToRetain.toList
+    }
+  }
+
+  private def cleanupCompactedFiles(files: Seq[FileStatus]): Unit = {
+    files.foreach { file =>
+      var deleted = false
+      try {
+        deleted = fs.delete(file.getPath, true)
+      } catch {
+        case _: IOException =>
+      }
+      if (!deleted) {
+        logWarning(s"Failed to remove ${file.getPath} / skip removing.")
+      }
+    }
+  }
+
+  private def findFilesToCompact(
+      eventLogFiles: Seq[FileStatus]): (Seq[FileStatus], Seq[FileStatus]) = {
+    val files = 
RollingEventLogFilesFileReader.dropBeforeLastCompactFile(eventLogFiles)
 
 Review comment:
   Not all app logs are going through compactor before reading; 
`RollingEventLogFilesFileReader` still has to provide such functionality. Maybe 
it would be better to have a prerequisite of parameter for `compact` (compact 
file should be only at most one and placed to the first if it has) and simply 
assert with `require`.

----------------------------------------------------------------
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:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to