corecursion commented on code in PR #15015: URL: https://github.com/apache/lucene/pull/15015#discussion_r2263039588
########## lucene/core/src/java/org/apache/lucene/index/MultiIndexMergeScheduler.java: ########## @@ -0,0 +1,206 @@ +/* + * 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.lucene.index; + +import java.io.IOException; +import org.apache.lucene.store.Directory; +import org.apache.lucene.util.InfoStream; + +/** + * Multi-index or multi-tenant merge scheduling. + * + * <p>MultiIndexMergeScheduler builds on existing functionality in ConcurrentMergeScheduler by + * automatically tracking merge sources and merge threads and the index Directory that each applies + * to, and then shunting all of them into a single ConcurrentMergeScheduler instance. + * + * <p>The multi-tenant merge scheduling can be used easily by creating a MultiIndexMergeScheduler + * instance for each index and then using each instance normally, the same way you would use a lone + * ConcurrentMergeScheduler. + * + * @lucene.experimental + */ +class MultiIndexMergeScheduler extends MergeScheduler { + private final Directory directory; + private final CombinedMergeScheduler combinedMergeScheduler; + private final boolean manageSingleton; + + /** The main MultiIndexMergeScheduler constructor -- use this one. */ + public MultiIndexMergeScheduler(Directory directory) { + this.directory = directory; + this.combinedMergeScheduler = CombinedMergeScheduler.incrementSingletonReference(); + this.manageSingleton = true; + } + + /** Alternate MultiIndexMergeScheduler constructor for unit testing or tenant grouping. */ + public MultiIndexMergeScheduler( + Directory directory, CombinedMergeScheduler combinedMergeScheduler) { + this.directory = directory; + this.combinedMergeScheduler = combinedMergeScheduler; + this.manageSingleton = false; + } + + public Directory getDirectory() { + return directory; + } + + public CombinedMergeScheduler getCombinedMergeScheduler() { + return combinedMergeScheduler; + } + + @Override + public void merge(MergeSource mergeSource, MergeTrigger trigger) throws IOException { + var taggedMergeSource = + new CombinedMergeScheduler.TaggedMergeSource(mergeSource, this.directory); + this.combinedMergeScheduler.merge(taggedMergeSource, trigger); + } + + @Override + public Directory wrapForMerge(MergePolicy.OneMerge merge, Directory in) { + return this.combinedMergeScheduler.wrapForMerge(merge, in); + } + + /** Close this scheduler for one directory/index. Called automatically by IndexWriter. */ + @Override + public void close() throws IOException { + this.combinedMergeScheduler.sync(this.directory); + if (this.manageSingleton) { + CombinedMergeScheduler.decrementSingletonReference(); + } + } + + // We created this method because we cannot easily override the initialize() method + // in ConcurrentMergeScheduler. We don't need the initDynamicDefaults() part in the + // initialize() method, and only need the setInfoStream(). + public void setInfoStream(InfoStream infoStream) { + this.combinedMergeScheduler.setInfoStream(infoStream); + } + + /** + * CombinedMergeScheduler is used internally by MultiIndexMergeScheduler to balance resources + * across multiple indices. Normally you don't need to use this. + * + * <p>For testing purposes, or if partitioning of tenants into groups is needed for some reason, a + * CombinedMergeScheduler can be provided to the MultiIndexMergeScheduler constructor. + * + * <p>CombinedMergeScheduler should <b><i>not</i></b> be passed directly to IndexWriter. + */ + static class CombinedMergeScheduler extends ConcurrentMergeScheduler { + private static CombinedMergeScheduler singleton = null; + private static int singletonRefCount = 0; + + private static synchronized CombinedMergeScheduler incrementSingletonReference() { + if (singleton == null) { + singleton = new CombinedMergeScheduler(); + } + singletonRefCount++; + return singleton; + } + + private static synchronized void decrementSingletonReference() throws IOException { + if (singletonRefCount < 1) { + throw new IllegalStateException("decrementSingletonReference() called too many times"); + } + singletonRefCount--; + if (singletonRefCount == 0) { + singleton = null; Review Comment: Ha, yes, calling `close()` was the point of the reference counting, wasn't it? Fixed in: [d4eec7a](https://github.com/apache/lucene/pull/15015/commits/d4eec7a7a4ad66ddac4c27a5a3325a028075272a) (The `close()` got lost when the code was switched from atomics to synchronized.) -- 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: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
