001/*
002 * Licensed to the Apache Software Foundation (ASF) under one
003 * or more contributor license agreements.  See the NOTICE file
004 * distributed with this work for additional information
005 * regarding copyright ownership.  The ASF licenses this file
006 * to you under the Apache License, Version 2.0 (the
007 * "License"); you may not use this file except in compliance
008 * with the License.  You may obtain a copy of the License at
009 *
010 *     http://www.apache.org/licenses/LICENSE-2.0
011 *
012 * Unless required by applicable law or agreed to in writing, software
013 * distributed under the License is distributed on an "AS IS" BASIS,
014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
015 * See the License for the specific language governing permissions and
016 * limitations under the License.
017 */
018package org.apache.hadoop.hbase.regionserver.storefiletracker;
019
020import java.io.IOException;
021import java.util.Collection;
022import java.util.List;
023import org.apache.hadoop.fs.FileStatus;
024import org.apache.hadoop.fs.Path;
025import org.apache.hadoop.hbase.TableName;
026import org.apache.hadoop.hbase.client.TableDescriptorBuilder;
027import org.apache.hadoop.hbase.io.Reference;
028import org.apache.hadoop.hbase.regionserver.CreateStoreFileWriterParams;
029import org.apache.hadoop.hbase.regionserver.HStoreFile;
030import org.apache.hadoop.hbase.regionserver.StoreFileInfo;
031import org.apache.hadoop.hbase.regionserver.StoreFileWriter;
032import org.apache.yetus.audience.InterfaceAudience;
033
034/**
035 * An interface to define how we track the store files for a give store.
036 * <p/>
037 * In the old time, we will write store to a tmp directory first, and then rename it to the actual
038 * data file. And once a store file is under data directory, we will consider it as 'committed'. And
039 * we need to do listing when loading store files.
040 * <p/>
041 * When cloud age is coming, now we want to store the store files on object storage, where rename
042 * and list are not as cheap as on HDFS, especially rename. Although introducing a metadata
043 * management layer for object storage could solve the problem, but we still want HBase to run on
044 * pure object storage, so here we introduce this interface to abstract how we track the store
045 * files. For the old implementation, we just persist nothing here, and do listing to load store
046 * files. When running on object storage, we could persist the store file list in a system region,
047 * or in a file on the object storage, to make it possible to write directly into the data directory
048 * to avoid renaming, and also avoid listing when loading store files.
049 * <p/>
050 * The implementation requires to be thread safe as flush and compaction may occur as the same time,
051 * and we could also do multiple compactions at the same time. As the implementation may choose to
052 * persist the store file list to external storage, which could be slow, it is the duty for the
053 * callers to not call it inside a lock which may block normal read/write requests.
054 */
055@InterfaceAudience.Private
056public interface StoreFileTracker {
057  /**
058   * Load the store files list when opening a region.
059   */
060  List<StoreFileInfo> load() throws IOException;
061
062  /**
063   * Add new store files.
064   * <p/>
065   * Used for flush and bulk load.
066   */
067  void add(Collection<StoreFileInfo> newFiles) throws IOException;
068
069  /**
070   * Add new store files and remove compacted store files after compaction.
071   */
072  void replace(Collection<StoreFileInfo> compactedFiles, Collection<StoreFileInfo> newFiles)
073    throws IOException;
074
075  /**
076   * Set the store files.
077   */
078  void set(List<StoreFileInfo> files) throws IOException;
079
080  /**
081   * Create a writer for writing new store files.
082   * @return Writer for a new StoreFile
083   */
084  StoreFileWriter createWriter(CreateStoreFileWriterParams params) throws IOException;
085
086  /**
087   * Adds StoreFileTracker implementations specific configurations into the table descriptor.
088   * <p/>
089   * This is used to avoid accidentally data loss when changing the cluster level store file tracker
090   * implementation, and also possible misconfiguration between master and region servers.
091   * <p/>
092   * See HBASE-26246 for more details.
093   * @param builder The table descriptor builder for the given table.
094   */
095  TableDescriptorBuilder updateWithTrackerConfigs(TableDescriptorBuilder builder);
096
097  /**
098   * Whether the implementation of this tracker requires you to write to temp directory first, i.e,
099   * does not allow broken store files under the actual data directory.
100   */
101  boolean requireWritingToTmpDirFirst();
102
103  Reference createReference(Reference reference, Path path) throws IOException;
104
105  /**
106   * Reads the reference file from the given path.
107   * @param path the {@link Path} to the reference file in the file system.
108   * @return a {@link Reference} that points at top/bottom half of a an hfile
109   */
110  Reference readReference(Path path) throws IOException;
111
112  /**
113   * Returns true if the specified family has reference files
114   * @return true if family contains reference files
115   */
116  boolean hasReferences() throws IOException;
117
118  StoreFileInfo getStoreFileInfo(final FileStatus fileStatus, final Path initialPath,
119    final boolean primaryReplica) throws IOException;
120
121  StoreFileInfo getStoreFileInfo(final Path initialPath, final boolean primaryReplica)
122    throws IOException;
123
124  /**
125   * Create a new HFileLink
126   * <p>
127   * It also adds a back-reference to the hfile back-reference directory to simplify the
128   * reference-count and the cleaning process.
129   * @param hfileLinkName - HFileLink name (it contains hfile-region-table)
130   * @param createBackRef - Whether back reference should be created. Defaults to true.
131   * @return the file link name.
132   * @throws IOException on file or parent directory creation failure.
133   */
134  String createHFileLink(final TableName linkedTable, final String linkedRegion,
135    final String hfileName, final boolean createBackRef) throws IOException;
136
137  /**
138   * Create a new HFileLink starting from a hfileLink name
139   * <p>
140   * It also adds a back-reference to the hfile back-reference directory to simplify the
141   * reference-count and the cleaning process.
142   * @param hfileLinkName - HFileLink name (it contains hfile-region-table)
143   * @param createBackRef - Whether back reference should be created. Defaults to true.
144   * @return the file link name.
145   * @throws IOException on file or parent directory creation failure.
146   */
147  String createFromHFileLink(final String hfileName, final boolean createBackRef)
148    throws IOException;
149
150  /**
151   * Closes and archives the specified store files from the specified family.
152   * @param storeFiles set of store files to remove
153   * @throws IOException if the archiving fails
154   */
155  void removeStoreFiles(List<HStoreFile> storeFiles) throws IOException;
156}