View Javadoc

1   /**
2    * Licensed to the Apache Software Foundation (ASF) under one
3    * or more contributor license agreements.  See the NOTICE file
4    * distributed with this work for additional information
5    * regarding copyright ownership.  The ASF licenses this file
6    * to you under the Apache License, Version 2.0 (the
7    * "License"); you may not use this file except in compliance
8    * with the License.  You may obtain a copy of the License at
9    *
10   *     http://www.apache.org/licenses/LICENSE-2.0
11   *
12   * Unless required by applicable law or agreed to in writing, software
13   * distributed under the License is distributed on an "AS IS" BASIS,
14   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15   * See the License for the specific language governing permissions and
16   * limitations under the License.
17   */
18  package org.apache.hadoop.hbase.regionserver;
19  
20  import java.io.FileNotFoundException;
21  import java.io.IOException;
22  import java.util.Collection;
23  import java.util.List;
24  import java.util.NavigableSet;
25  
26  import org.apache.hadoop.fs.FileSystem;
27  import org.apache.hadoop.fs.Path;
28  import org.apache.hadoop.hbase.Cell;
29  import org.apache.hadoop.hbase.HColumnDescriptor;
30  import org.apache.hadoop.hbase.HRegionInfo;
31  import org.apache.hadoop.hbase.KeyValue;
32  import org.apache.hadoop.hbase.TableName;
33  import org.apache.hadoop.hbase.classification.InterfaceAudience;
34  import org.apache.hadoop.hbase.classification.InterfaceStability;
35  import org.apache.hadoop.hbase.client.Scan;
36  import org.apache.hadoop.hbase.io.HeapSize;
37  import org.apache.hadoop.hbase.io.compress.Compression;
38  import org.apache.hadoop.hbase.io.hfile.CacheConfig;
39  import org.apache.hadoop.hbase.io.hfile.HFileDataBlockEncoder;
40  import org.apache.hadoop.hbase.protobuf.generated.WALProtos.CompactionDescriptor;
41  import org.apache.hadoop.hbase.regionserver.compactions.CompactionContext;
42  import org.apache.hadoop.hbase.regionserver.compactions.CompactionProgress;
43  import org.apache.hadoop.hbase.regionserver.compactions.CompactionRequest;
44  import org.apache.hadoop.hbase.regionserver.compactions.CompactionThroughputController;
45  
46  /**
47   * Interface for objects that hold a column family in a Region. Its a memstore and a set of zero or
48   * more StoreFiles, which stretch backwards over time.
49   */
50  @InterfaceAudience.Private
51  @InterfaceStability.Evolving
52  public interface Store extends HeapSize, StoreConfigInformation {
53  
54    /* The default priority for user-specified compaction requests.
55     * The user gets top priority unless we have blocking compactions. (Pri <= 0)
56     */ int PRIORITY_USER = 1;
57    int NO_PRIORITY = Integer.MIN_VALUE;
58  
59    // General Accessors
60    KeyValue.KVComparator getComparator();
61  
62    Collection<StoreFile> getStorefiles();
63  
64    /**
65     * Close all the readers We don't need to worry about subsequent requests because the HRegion
66     * holds a write lock that will prevent any more reads or writes.
67     * @return the {@link StoreFile StoreFiles} that were previously being used.
68     * @throws IOException on failure
69     */
70    Collection<StoreFile> close() throws IOException;
71  
72    /**
73     * Return a scanner for both the memstore and the HStore files. Assumes we are not in a
74     * compaction.
75     * @param scan Scan to apply when scanning the stores
76     * @param targetCols columns to scan
77     * @return a scanner over the current key values
78     * @throws IOException on failure
79     */
80    KeyValueScanner getScanner(Scan scan, final NavigableSet<byte[]> targetCols, long readPt)
81        throws IOException;
82  
83    /**
84     * Get all scanners with no filtering based on TTL (that happens further down
85     * the line).
86     * @param cacheBlocks
87     * @param isGet
88     * @param usePread
89     * @param isCompaction
90     * @param matcher
91     * @param startRow
92     * @param stopRow
93     * @param readPt
94     * @return all scanners for this store
95     */
96    List<KeyValueScanner> getScanners(
97      boolean cacheBlocks,
98      boolean isGet,
99      boolean usePread,
100     boolean isCompaction,
101     ScanQueryMatcher matcher,
102     byte[] startRow,
103     byte[] stopRow,
104     long readPt
105   ) throws IOException;
106 
107   ScanInfo getScanInfo();
108 
109   /**
110    * Adds or replaces the specified KeyValues.
111    * <p>
112    * For each KeyValue specified, if a cell with the same row, family, and qualifier exists in
113    * MemStore, it will be replaced. Otherwise, it will just be inserted to MemStore.
114    * <p>
115    * This operation is atomic on each KeyValue (row/family/qualifier) but not necessarily atomic
116    * across all of them.
117    * @param cells
118    * @param readpoint readpoint below which we can safely remove duplicate KVs
119    * @return memstore size delta
120    * @throws IOException
121    */
122   long upsert(Iterable<Cell> cells, long readpoint) throws IOException;
123 
124   /**
125    * Adds a value to the memstore
126    * @param kv
127    * @return memstore size delta
128    */
129   long add(KeyValue kv);
130 
131   /**
132    * When was the last edit done in the memstore
133    */
134   long timeOfOldestEdit();
135 
136   /**
137    * Removes a kv from the memstore. The KeyValue is removed only if its key & memstoreTS match the
138    * key & memstoreTS value of the kv parameter.
139    * @param kv
140    */
141   void rollback(final KeyValue kv);
142 
143   /**
144    * Find the key that matches <i>row</i> exactly, or the one that immediately precedes it. WARNING:
145    * Only use this method on a table where writes occur with strictly increasing timestamps. This
146    * method assumes this pattern of writes in order to make it reasonably performant. Also our
147    * search is dependent on the axiom that deletes are for cells that are in the container that
148    * follows whether a memstore snapshot or a storefile, not for the current container: i.e. we'll
149    * see deletes before we come across cells we are to delete. Presumption is that the
150    * memstore#kvset is processed before memstore#snapshot and so on.
151    * @param row The row key of the targeted row.
152    * @return Found keyvalue or null if none found.
153    * @throws IOException
154    */
155   KeyValue getRowKeyAtOrBefore(final byte[] row) throws IOException;
156 
157   FileSystem getFileSystem();
158 
159   /*
160    * @param maxKeyCount
161    * @param compression Compression algorithm to use
162    * @param isCompaction whether we are creating a new file in a compaction
163    * @param includeMVCCReadpoint whether we should out the MVCC readpoint
164    * @return Writer for a new StoreFile in the tmp dir.
165    */
166   StoreFile.Writer createWriterInTmp(
167     long maxKeyCount,
168     Compression.Algorithm compression,
169     boolean isCompaction,
170     boolean includeMVCCReadpoint,
171     boolean includesTags
172   ) throws IOException;
173 
174   // Compaction oriented methods
175 
176   boolean throttleCompaction(long compactionSize);
177 
178   /**
179    * getter for CompactionProgress object
180    * @return CompactionProgress object; can be null
181    */
182   CompactionProgress getCompactionProgress();
183 
184   CompactionContext requestCompaction() throws IOException;
185 
186   CompactionContext requestCompaction(int priority, CompactionRequest baseRequest)
187       throws IOException;
188 
189   void cancelRequestedCompaction(CompactionContext compaction);
190 
191   List<StoreFile> compact(CompactionContext compaction,
192       CompactionThroughputController throughputController) throws IOException;
193 
194   /**
195    * @return true if we should run a major compaction.
196    */
197   boolean isMajorCompaction() throws IOException;
198 
199   void triggerMajorCompaction();
200 
201   /**
202    * See if there's too much store files in this store
203    * @return true if number of store files is greater than the number defined in minFilesToCompact
204    */
205   boolean needsCompaction();
206 
207   int getCompactPriority();
208 
209   StoreFlushContext createFlushContext(long cacheFlushId);
210 
211   /**
212    * Call to complete a compaction. Its for the case where we find in the WAL a compaction
213    * that was not finished.  We could find one recovering a WAL after a regionserver crash.
214    * See HBASE-2331.
215    * @param compaction
216    */
217   void completeCompactionMarker(CompactionDescriptor compaction)
218       throws IOException;
219 
220   // Split oriented methods
221 
222   boolean canSplit();
223 
224   /**
225    * Determines if Store should be split
226    * @return byte[] if store should be split, null otherwise.
227    */
228   byte[] getSplitPoint();
229 
230   // Bulk Load methods
231 
232   /**
233    * This throws a WrongRegionException if the HFile does not fit in this region, or an
234    * InvalidHFileException if the HFile is not valid.
235    */
236   void assertBulkLoadHFileOk(Path srcPath) throws IOException;
237 
238   /**
239    * This method should only be called from HRegion. It is assumed that the ranges of values in the
240    * HFile fit within the stores assigned region. (assertBulkLoadHFileOk checks this)
241    *
242    * @param srcPathStr
243    * @param sequenceId sequence Id associated with the HFile
244    */
245   void bulkLoadHFile(String srcPathStr, long sequenceId) throws IOException;
246 
247   // General accessors into the state of the store
248   // TODO abstract some of this out into a metrics class
249 
250   /**
251    * @return <tt>true</tt> if the store has any underlying reference files to older HFiles
252    */
253   boolean hasReferences();
254 
255   /**
256    * @return The size of this store's memstore, in bytes
257    */
258   long getMemStoreSize();
259 
260   /**
261    * @return The amount of memory we could flush from this memstore; usually this is equal to
262    * {@link #getMemStoreSize()} unless we are carrying snapshots and then it will be the size of
263    * outstanding snapshots.
264    */
265   long getFlushableSize();
266 
267   HColumnDescriptor getFamily();
268 
269   /**
270    * @return The maximum memstoreTS in all store files.
271    */
272   long getMaxMemstoreTS();
273 
274   /**
275    * @return the data block encoder
276    */
277   HFileDataBlockEncoder getDataBlockEncoder();
278 
279   /** @return aggregate size of all HStores used in the last compaction */
280   long getLastCompactSize();
281 
282   /** @return aggregate size of HStore */
283   long getSize();
284 
285   /**
286    * @return Count of store files
287    */
288   int getStorefilesCount();
289 
290   /**
291    * @return The size of the store files, in bytes, uncompressed.
292    */
293   long getStoreSizeUncompressed();
294 
295   /**
296    * @return The size of the store files, in bytes.
297    */
298   long getStorefilesSize();
299 
300   /**
301    * @return The size of the store file indexes, in bytes.
302    */
303   long getStorefilesIndexSize();
304 
305   /**
306    * Returns the total size of all index blocks in the data block indexes, including the root level,
307    * intermediate levels, and the leaf level for multi-level indexes, or just the root level for
308    * single-level indexes.
309    * @return the total size of block indexes in the store
310    */
311   long getTotalStaticIndexSize();
312 
313   /**
314    * Returns the total byte size of all Bloom filter bit arrays. For compound Bloom filters even the
315    * Bloom blocks currently not loaded into the block cache are counted.
316    * @return the total size of all Bloom filters in the store
317    */
318   long getTotalStaticBloomSize();
319 
320   // Test-helper methods
321 
322   /**
323    * Used for tests.
324    * @return cache configuration for this Store.
325    */
326   CacheConfig getCacheConfig();
327 
328   /**
329    * @return the parent region info hosting this store
330    */
331   HRegionInfo getRegionInfo();
332 
333   RegionCoprocessorHost getCoprocessorHost();
334 
335   boolean areWritesEnabled();
336 
337   /**
338    * @return The smallest mvcc readPoint across all the scanners in this
339    * region. Writes older than this readPoint, are included  in every
340    * read operation.
341    */
342   long getSmallestReadPoint();
343 
344   String getColumnFamilyName();
345 
346   TableName getTableName();
347 
348   /**
349    * @return The number of cells flushed to disk
350    */
351   long getFlushedCellsCount();
352 
353   /**
354    * @return The total size of data flushed to disk, in bytes
355    */
356   long getFlushedCellsSize();
357 
358   /**
359    * @return The number of cells processed during minor compactions
360    */
361   long getCompactedCellsCount();
362 
363   /**
364    * @return The total amount of data processed during minor compactions, in bytes
365    */
366   long getCompactedCellsSize();
367 
368   /**
369    * @return The number of cells processed during major compactions
370    */
371   long getMajorCompactedCellsCount();
372 
373   /**
374    * @return The total amount of data processed during major compactions, in bytes
375    */
376   long getMajorCompactedCellsSize();
377 
378   /*
379    * @param o Observer who wants to know about changes in set of Readers
380    */
381   void addChangedReaderObserver(ChangedReadersObserver o);
382 
383   /*
384    * @param o Observer no longer interested in changes in set of Readers.
385    */
386   void deleteChangedReaderObserver(ChangedReadersObserver o);
387 
388   /**
389    * @return Whether this store has too many store files.
390    */
391   boolean hasTooManyStoreFiles();
392 
393   /**
394    * This value can represent the degree of emergency of compaction for this store. It should be
395    * greater than or equal to 0.0, any value greater than 1.0 means we have too many store files.
396    * <ul>
397    * <li>if getStorefilesCount &lt;= getMinFilesToCompact, return 0.0</li>
398    * <li>return (getStorefilesCount - getMinFilesToCompact) / (blockingFileCount -
399    * getMinFilesToCompact)</li>
400    * </ul>
401    * <p>
402    * And for striped stores, we should calculate this value by the files in each stripe separately
403    * and return the maximum value.
404    * <p>
405    * It is similar to {@link #getCompactPriority()} except that it is more suitable to use in a
406    * linear formula.
407    */
408   double getCompactionPressure();
409 }