http://git-wip-us.apache.org/repos/asf/hbase-site/blob/a7f7dba1/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html ---------------------------------------------------------------------- diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html b/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html index 99f53c4..2d5c946 100644 --- a/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html +++ b/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html @@ -123,7 +123,7 @@ <span class="sourceLineNo">115</span> public static final boolean SPLIT_SKIP_ERRORS_DEFAULT = false;<a name="line.115"></a> <span class="sourceLineNo">116</span><a name="line.116"></a> <span class="sourceLineNo">117</span> // Parameters for split process<a name="line.117"></a> -<span class="sourceLineNo">118</span> protected final Path rootDir;<a name="line.118"></a> +<span class="sourceLineNo">118</span> protected final Path walDir;<a name="line.118"></a> <span class="sourceLineNo">119</span> protected final FileSystem fs;<a name="line.119"></a> <span class="sourceLineNo">120</span> protected final Configuration conf;<a name="line.120"></a> <span class="sourceLineNo">121</span><a name="line.121"></a> @@ -156,14 +156,14 @@ <span class="sourceLineNo">148</span><a name="line.148"></a> <span class="sourceLineNo">149</span><a name="line.149"></a> <span class="sourceLineNo">150</span> @VisibleForTesting<a name="line.150"></a> -<span class="sourceLineNo">151</span> WALSplitter(final WALFactory factory, Configuration conf, Path rootDir,<a name="line.151"></a> +<span class="sourceLineNo">151</span> WALSplitter(final WALFactory factory, Configuration conf, Path walDir,<a name="line.151"></a> <span class="sourceLineNo">152</span> FileSystem fs, LastSequenceId idChecker,<a name="line.152"></a> <span class="sourceLineNo">153</span> SplitLogWorkerCoordination splitLogWorkerCoordination) {<a name="line.153"></a> <span class="sourceLineNo">154</span> this.conf = HBaseConfiguration.create(conf);<a name="line.154"></a> <span class="sourceLineNo">155</span> String codecClassName = conf<a name="line.155"></a> <span class="sourceLineNo">156</span> .get(WALCellCodec.WAL_CELL_CODEC_CLASS_KEY, WALCellCodec.class.getName());<a name="line.156"></a> <span class="sourceLineNo">157</span> this.conf.set(HConstants.RPC_CODEC_CONF_KEY, codecClassName);<a name="line.157"></a> -<span class="sourceLineNo">158</span> this.rootDir = rootDir;<a name="line.158"></a> +<span class="sourceLineNo">158</span> this.walDir = walDir;<a name="line.158"></a> <span class="sourceLineNo">159</span> this.fs = fs;<a name="line.159"></a> <span class="sourceLineNo">160</span> this.sequenceIdChecker = idChecker;<a name="line.160"></a> <span class="sourceLineNo">161</span> this.splitLogWorkerCoordination = splitLogWorkerCoordination;<a name="line.161"></a> @@ -194,11 +194,11 @@ <span class="sourceLineNo">186</span> * <p><a name="line.186"></a> <span class="sourceLineNo">187</span> * @return false if it is interrupted by the progress-able.<a name="line.187"></a> <span class="sourceLineNo">188</span> */<a name="line.188"></a> -<span class="sourceLineNo">189</span> public static boolean splitLogFile(Path rootDir, FileStatus logfile, FileSystem fs,<a name="line.189"></a> +<span class="sourceLineNo">189</span> public static boolean splitLogFile(Path walDir, FileStatus logfile, FileSystem fs,<a name="line.189"></a> <span class="sourceLineNo">190</span> Configuration conf, CancelableProgressable reporter, LastSequenceId idChecker,<a name="line.190"></a> <span class="sourceLineNo">191</span> SplitLogWorkerCoordination splitLogWorkerCoordination, final WALFactory factory)<a name="line.191"></a> <span class="sourceLineNo">192</span> throws IOException {<a name="line.192"></a> -<span class="sourceLineNo">193</span> WALSplitter s = new WALSplitter(factory, conf, rootDir, fs, idChecker,<a name="line.193"></a> +<span class="sourceLineNo">193</span> WALSplitter s = new WALSplitter(factory, conf, walDir, fs, idChecker,<a name="line.193"></a> <span class="sourceLineNo">194</span> splitLogWorkerCoordination);<a name="line.194"></a> <span class="sourceLineNo">195</span> return s.splitLogFile(logfile, reporter);<a name="line.195"></a> <span class="sourceLineNo">196</span> }<a name="line.196"></a> @@ -330,10 +330,10 @@ <span class="sourceLineNo">322</span> LOG.warn("Could not parse, corrupted WAL={}", logPath, e);<a name="line.322"></a> <span class="sourceLineNo">323</span> if (splitLogWorkerCoordination != null) {<a name="line.323"></a> <span class="sourceLineNo">324</span> // Some tests pass in a csm of null.<a name="line.324"></a> -<span class="sourceLineNo">325</span> splitLogWorkerCoordination.markCorrupted(rootDir, logfile.getPath().getName(), fs);<a name="line.325"></a> +<span class="sourceLineNo">325</span> splitLogWorkerCoordination.markCorrupted(walDir, logfile.getPath().getName(), fs);<a name="line.325"></a> <span class="sourceLineNo">326</span> } else {<a name="line.326"></a> <span class="sourceLineNo">327</span> // for tests only<a name="line.327"></a> -<span class="sourceLineNo">328</span> ZKSplitLog.markCorrupted(rootDir, logfile.getPath().getName(), fs);<a name="line.328"></a> +<span class="sourceLineNo">328</span> ZKSplitLog.markCorrupted(walDir, logfile.getPath().getName(), fs);<a name="line.328"></a> <span class="sourceLineNo">329</span> }<a name="line.329"></a> <span class="sourceLineNo">330</span> isCorrupted = true;<a name="line.330"></a> <span class="sourceLineNo">331</span> } catch (IOException e) {<a name="line.331"></a> @@ -465,1437 +465,1442 @@ <span class="sourceLineNo">457</span> * <code>logEntry</code>: e.g. /hbase/some_table/2323432434/recovered.edits/2332.<a name="line.457"></a> <span class="sourceLineNo">458</span> * This method also ensures existence of RECOVERED_EDITS_DIR under the region<a name="line.458"></a> <span class="sourceLineNo">459</span> * creating it if necessary.<a name="line.459"></a> -<span class="sourceLineNo">460</span> * @param fs<a name="line.460"></a> -<span class="sourceLineNo">461</span> * @param logEntry<a name="line.461"></a> -<span class="sourceLineNo">462</span> * @param rootDir HBase root dir.<a name="line.462"></a> -<span class="sourceLineNo">463</span> * @param fileNameBeingSplit the file being split currently. Used to generate tmp file name.<a name="line.463"></a> -<span class="sourceLineNo">464</span> * @return Path to file into which to dump split log edits.<a name="line.464"></a> -<span class="sourceLineNo">465</span> * @throws IOException<a name="line.465"></a> -<span class="sourceLineNo">466</span> */<a name="line.466"></a> -<span class="sourceLineNo">467</span> @SuppressWarnings("deprecation")<a name="line.467"></a> -<span class="sourceLineNo">468</span> @VisibleForTesting<a name="line.468"></a> -<span class="sourceLineNo">469</span> static Path getRegionSplitEditsPath(final FileSystem fs,<a name="line.469"></a> -<span class="sourceLineNo">470</span> final Entry logEntry, final Path rootDir, String fileNameBeingSplit)<a name="line.470"></a> -<span class="sourceLineNo">471</span> throws IOException {<a name="line.471"></a> -<span class="sourceLineNo">472</span> Path tableDir = FSUtils.getTableDir(rootDir, logEntry.getKey().getTableName());<a name="line.472"></a> -<span class="sourceLineNo">473</span> String encodedRegionName = Bytes.toString(logEntry.getKey().getEncodedRegionName());<a name="line.473"></a> -<span class="sourceLineNo">474</span> Path regiondir = HRegion.getRegionDir(tableDir, encodedRegionName);<a name="line.474"></a> -<span class="sourceLineNo">475</span> Path dir = getRegionDirRecoveredEditsDir(regiondir);<a name="line.475"></a> -<span class="sourceLineNo">476</span><a name="line.476"></a> -<span class="sourceLineNo">477</span> if (!fs.exists(regiondir)) {<a name="line.477"></a> -<span class="sourceLineNo">478</span> LOG.info("This region's directory does not exist: {}."<a name="line.478"></a> -<span class="sourceLineNo">479</span> + "It is very likely that it was already split so it is "<a name="line.479"></a> -<span class="sourceLineNo">480</span> + "safe to discard those edits.", regiondir);<a name="line.480"></a> -<span class="sourceLineNo">481</span> return null;<a name="line.481"></a> -<span class="sourceLineNo">482</span> }<a name="line.482"></a> -<span class="sourceLineNo">483</span> if (fs.exists(dir) && fs.isFile(dir)) {<a name="line.483"></a> -<span class="sourceLineNo">484</span> Path tmp = new Path("/tmp");<a name="line.484"></a> -<span class="sourceLineNo">485</span> if (!fs.exists(tmp)) {<a name="line.485"></a> -<span class="sourceLineNo">486</span> fs.mkdirs(tmp);<a name="line.486"></a> -<span class="sourceLineNo">487</span> }<a name="line.487"></a> -<span class="sourceLineNo">488</span> tmp = new Path(tmp,<a name="line.488"></a> -<span class="sourceLineNo">489</span> HConstants.RECOVERED_EDITS_DIR + "_" + encodedRegionName);<a name="line.489"></a> -<span class="sourceLineNo">490</span> LOG.warn("Found existing old file: {}. It could be some "<a name="line.490"></a> -<span class="sourceLineNo">491</span> + "leftover of an old installation. It should be a folder instead. "<a name="line.491"></a> -<span class="sourceLineNo">492</span> + "So moving it to {}", dir, tmp);<a name="line.492"></a> -<span class="sourceLineNo">493</span> if (!fs.rename(dir, tmp)) {<a name="line.493"></a> -<span class="sourceLineNo">494</span> LOG.warn("Failed to sideline old file {}", dir);<a name="line.494"></a> -<span class="sourceLineNo">495</span> }<a name="line.495"></a> -<span class="sourceLineNo">496</span> }<a name="line.496"></a> -<span class="sourceLineNo">497</span><a name="line.497"></a> -<span class="sourceLineNo">498</span> if (!fs.exists(dir) && !fs.mkdirs(dir)) {<a name="line.498"></a> -<span class="sourceLineNo">499</span> LOG.warn("mkdir failed on {}", dir);<a name="line.499"></a> -<span class="sourceLineNo">500</span> }<a name="line.500"></a> -<span class="sourceLineNo">501</span> // Append fileBeingSplit to prevent name conflict since we may have duplicate wal entries now.<a name="line.501"></a> -<span class="sourceLineNo">502</span> // Append file name ends with RECOVERED_LOG_TMPFILE_SUFFIX to ensure<a name="line.502"></a> -<span class="sourceLineNo">503</span> // region's replayRecoveredEdits will not delete it<a name="line.503"></a> -<span class="sourceLineNo">504</span> String fileName = formatRecoveredEditsFileName(logEntry.getKey().getSequenceId());<a name="line.504"></a> -<span class="sourceLineNo">505</span> fileName = getTmpRecoveredEditsFileName(fileName + "-" + fileNameBeingSplit);<a name="line.505"></a> -<span class="sourceLineNo">506</span> return new Path(dir, fileName);<a name="line.506"></a> -<span class="sourceLineNo">507</span> }<a name="line.507"></a> -<span class="sourceLineNo">508</span><a name="line.508"></a> -<span class="sourceLineNo">509</span> private static String getTmpRecoveredEditsFileName(String fileName) {<a name="line.509"></a> -<span class="sourceLineNo">510</span> return fileName + RECOVERED_LOG_TMPFILE_SUFFIX;<a name="line.510"></a> -<span class="sourceLineNo">511</span> }<a name="line.511"></a> -<span class="sourceLineNo">512</span><a name="line.512"></a> -<span class="sourceLineNo">513</span> /**<a name="line.513"></a> -<span class="sourceLineNo">514</span> * Get the completed recovered edits file path, renaming it to be by last edit<a name="line.514"></a> -<span class="sourceLineNo">515</span> * in the file from its first edit. Then we could use the name to skip<a name="line.515"></a> -<span class="sourceLineNo">516</span> * recovered edits when doing {@link HRegion#replayRecoveredEditsIfAny}.<a name="line.516"></a> -<span class="sourceLineNo">517</span> * @param srcPath<a name="line.517"></a> -<span class="sourceLineNo">518</span> * @param maximumEditLogSeqNum<a name="line.518"></a> -<span class="sourceLineNo">519</span> * @return dstPath take file's last edit log seq num as the name<a name="line.519"></a> -<span class="sourceLineNo">520</span> */<a name="line.520"></a> -<span class="sourceLineNo">521</span> private static Path getCompletedRecoveredEditsFilePath(Path srcPath,<a name="line.521"></a> -<span class="sourceLineNo">522</span> long maximumEditLogSeqNum) {<a name="line.522"></a> -<span class="sourceLineNo">523</span> String fileName = formatRecoveredEditsFileName(maximumEditLogSeqNum);<a name="line.523"></a> -<span class="sourceLineNo">524</span> return new Path(srcPath.getParent(), fileName);<a name="line.524"></a> -<span class="sourceLineNo">525</span> }<a name="line.525"></a> -<span class="sourceLineNo">526</span><a name="line.526"></a> -<span class="sourceLineNo">527</span> @VisibleForTesting<a name="line.527"></a> -<span class="sourceLineNo">528</span> static String formatRecoveredEditsFileName(final long seqid) {<a name="line.528"></a> -<span class="sourceLineNo">529</span> return String.format("%019d", seqid);<a name="line.529"></a> -<span class="sourceLineNo">530</span> }<a name="line.530"></a> -<span class="sourceLineNo">531</span><a name="line.531"></a> -<span class="sourceLineNo">532</span> private static final Pattern EDITFILES_NAME_PATTERN = Pattern.compile("-?[0-9]+");<a name="line.532"></a> -<span class="sourceLineNo">533</span> private static final String RECOVERED_LOG_TMPFILE_SUFFIX = ".temp";<a name="line.533"></a> -<span class="sourceLineNo">534</span><a name="line.534"></a> -<span class="sourceLineNo">535</span> /**<a name="line.535"></a> -<span class="sourceLineNo">536</span> * @param regiondir<a name="line.536"></a> -<span class="sourceLineNo">537</span> * This regions directory in the filesystem.<a name="line.537"></a> -<span class="sourceLineNo">538</span> * @return The directory that holds recovered edits files for the region<a name="line.538"></a> -<span class="sourceLineNo">539</span> * <code>regiondir</code><a name="line.539"></a> -<span class="sourceLineNo">540</span> */<a name="line.540"></a> -<span class="sourceLineNo">541</span> public static Path getRegionDirRecoveredEditsDir(final Path regiondir) {<a name="line.541"></a> -<span class="sourceLineNo">542</span> return new Path(regiondir, HConstants.RECOVERED_EDITS_DIR);<a name="line.542"></a> -<span class="sourceLineNo">543</span> }<a name="line.543"></a> -<span class="sourceLineNo">544</span><a name="line.544"></a> -<span class="sourceLineNo">545</span> /**<a name="line.545"></a> -<span class="sourceLineNo">546</span> * Returns sorted set of edit files made by splitter, excluding files<a name="line.546"></a> -<span class="sourceLineNo">547</span> * with '.temp' suffix.<a name="line.547"></a> -<span class="sourceLineNo">548</span> *<a name="line.548"></a> -<span class="sourceLineNo">549</span> * @param fs<a name="line.549"></a> -<span class="sourceLineNo">550</span> * @param regiondir<a name="line.550"></a> -<span class="sourceLineNo">551</span> * @return Files in passed <code>regiondir</code> as a sorted set.<a name="line.551"></a> -<span class="sourceLineNo">552</span> * @throws IOException<a name="line.552"></a> -<span class="sourceLineNo">553</span> */<a name="line.553"></a> -<span class="sourceLineNo">554</span> public static NavigableSet<Path> getSplitEditFilesSorted(final FileSystem fs,<a name="line.554"></a> -<span class="sourceLineNo">555</span> final Path regiondir) throws IOException {<a name="line.555"></a> -<span class="sourceLineNo">556</span> NavigableSet<Path> filesSorted = new TreeSet<>();<a name="line.556"></a> -<span class="sourceLineNo">557</span> Path editsdir = getRegionDirRecoveredEditsDir(regiondir);<a name="line.557"></a> -<span class="sourceLineNo">558</span> if (!fs.exists(editsdir)) {<a name="line.558"></a> -<span class="sourceLineNo">559</span> return filesSorted;<a name="line.559"></a> -<span class="sourceLineNo">560</span> }<a name="line.560"></a> -<span class="sourceLineNo">561</span> FileStatus[] files = FSUtils.listStatus(fs, editsdir, new PathFilter() {<a name="line.561"></a> -<span class="sourceLineNo">562</span> @Override<a name="line.562"></a> -<span class="sourceLineNo">563</span> public boolean accept(Path p) {<a name="line.563"></a> -<span class="sourceLineNo">564</span> boolean result = false;<a name="line.564"></a> -<span class="sourceLineNo">565</span> try {<a name="line.565"></a> -<span class="sourceLineNo">566</span> // Return files and only files that match the editfile names pattern.<a name="line.566"></a> -<span class="sourceLineNo">567</span> // There can be other files in this directory other than edit files.<a name="line.567"></a> -<span class="sourceLineNo">568</span> // In particular, on error, we'll move aside the bad edit file giving<a name="line.568"></a> -<span class="sourceLineNo">569</span> // it a timestamp suffix. See moveAsideBadEditsFile.<a name="line.569"></a> -<span class="sourceLineNo">570</span> Matcher m = EDITFILES_NAME_PATTERN.matcher(p.getName());<a name="line.570"></a> -<span class="sourceLineNo">571</span> result = fs.isFile(p) && m.matches();<a name="line.571"></a> -<span class="sourceLineNo">572</span> // Skip the file whose name ends with RECOVERED_LOG_TMPFILE_SUFFIX,<a name="line.572"></a> -<span class="sourceLineNo">573</span> // because it means splitwal thread is writting this file.<a name="line.573"></a> -<span class="sourceLineNo">574</span> if (p.getName().endsWith(RECOVERED_LOG_TMPFILE_SUFFIX)) {<a name="line.574"></a> -<span class="sourceLineNo">575</span> result = false;<a name="line.575"></a> -<span class="sourceLineNo">576</span> }<a name="line.576"></a> -<span class="sourceLineNo">577</span> // Skip SeqId Files<a name="line.577"></a> -<span class="sourceLineNo">578</span> if (isSequenceIdFile(p)) {<a name="line.578"></a> -<span class="sourceLineNo">579</span> result = false;<a name="line.579"></a> -<span class="sourceLineNo">580</span> }<a name="line.580"></a> -<span class="sourceLineNo">581</span> } catch (IOException e) {<a name="line.581"></a> -<span class="sourceLineNo">582</span> LOG.warn("Failed isFile check on {}", p, e);<a name="line.582"></a> -<span class="sourceLineNo">583</span> }<a name="line.583"></a> -<span class="sourceLineNo">584</span> return result;<a name="line.584"></a> -<span class="sourceLineNo">585</span> }<a name="line.585"></a> -<span class="sourceLineNo">586</span> });<a name="line.586"></a> -<span class="sourceLineNo">587</span> if (ArrayUtils.isNotEmpty(files)) {<a name="line.587"></a> -<span class="sourceLineNo">588</span> Arrays.asList(files).forEach(status -> filesSorted.add(status.getPath()));<a name="line.588"></a> -<span class="sourceLineNo">589</span> }<a name="line.589"></a> -<span class="sourceLineNo">590</span> return filesSorted;<a name="line.590"></a> -<span class="sourceLineNo">591</span> }<a name="line.591"></a> -<span class="sourceLineNo">592</span><a name="line.592"></a> -<span class="sourceLineNo">593</span> /**<a name="line.593"></a> -<span class="sourceLineNo">594</span> * Move aside a bad edits file.<a name="line.594"></a> -<span class="sourceLineNo">595</span> *<a name="line.595"></a> -<span class="sourceLineNo">596</span> * @param fs<a name="line.596"></a> -<span class="sourceLineNo">597</span> * @param edits<a name="line.597"></a> -<span class="sourceLineNo">598</span> * Edits file to move aside.<a name="line.598"></a> -<span class="sourceLineNo">599</span> * @return The name of the moved aside file.<a name="line.599"></a> -<span class="sourceLineNo">600</span> * @throws IOException<a name="line.600"></a> -<span class="sourceLineNo">601</span> */<a name="line.601"></a> -<span class="sourceLineNo">602</span> public static Path moveAsideBadEditsFile(final FileSystem fs, final Path edits)<a name="line.602"></a> -<span class="sourceLineNo">603</span> throws IOException {<a name="line.603"></a> -<span class="sourceLineNo">604</span> Path moveAsideName = new Path(edits.getParent(), edits.getName() + "."<a name="line.604"></a> -<span class="sourceLineNo">605</span> + System.currentTimeMillis());<a name="line.605"></a> -<span class="sourceLineNo">606</span> if (!fs.rename(edits, moveAsideName)) {<a name="line.606"></a> -<span class="sourceLineNo">607</span> LOG.warn("Rename failed from {} to {}", edits, moveAsideName);<a name="line.607"></a> -<span class="sourceLineNo">608</span> }<a name="line.608"></a> -<span class="sourceLineNo">609</span> return moveAsideName;<a name="line.609"></a> -<span class="sourceLineNo">610</span> }<a name="line.610"></a> -<span class="sourceLineNo">611</span><a name="line.611"></a> -<span class="sourceLineNo">612</span> private static final String SEQUENCE_ID_FILE_SUFFIX = ".seqid";<a name="line.612"></a> -<span class="sourceLineNo">613</span> private static final String OLD_SEQUENCE_ID_FILE_SUFFIX = "_seqid";<a name="line.613"></a> -<span class="sourceLineNo">614</span> private static final int SEQUENCE_ID_FILE_SUFFIX_LENGTH = SEQUENCE_ID_FILE_SUFFIX.length();<a name="line.614"></a> -<span class="sourceLineNo">615</span><a name="line.615"></a> -<span class="sourceLineNo">616</span> /**<a name="line.616"></a> -<span class="sourceLineNo">617</span> * Is the given file a region open sequence id file.<a name="line.617"></a> -<span class="sourceLineNo">618</span> */<a name="line.618"></a> -<span class="sourceLineNo">619</span> @VisibleForTesting<a name="line.619"></a> -<span class="sourceLineNo">620</span> public static boolean isSequenceIdFile(final Path file) {<a name="line.620"></a> -<span class="sourceLineNo">621</span> return file.getName().endsWith(SEQUENCE_ID_FILE_SUFFIX)<a name="line.621"></a> -<span class="sourceLineNo">622</span> || file.getName().endsWith(OLD_SEQUENCE_ID_FILE_SUFFIX);<a name="line.622"></a> -<span class="sourceLineNo">623</span> }<a name="line.623"></a> -<span class="sourceLineNo">624</span><a name="line.624"></a> -<span class="sourceLineNo">625</span> private static FileStatus[] getSequenceIdFiles(FileSystem fs, Path regionDir) throws IOException {<a name="line.625"></a> -<span class="sourceLineNo">626</span> // TODO: Why are we using a method in here as part of our normal region open where<a name="line.626"></a> -<span class="sourceLineNo">627</span> // there is no splitting involved? Fix. St.Ack 01/20/2017.<a name="line.627"></a> -<span class="sourceLineNo">628</span> Path editsDir = WALSplitter.getRegionDirRecoveredEditsDir(regionDir);<a name="line.628"></a> -<span class="sourceLineNo">629</span> try {<a name="line.629"></a> -<span class="sourceLineNo">630</span> FileStatus[] files = fs.listStatus(editsDir, WALSplitter::isSequenceIdFile);<a name="line.630"></a> -<span class="sourceLineNo">631</span> return files != null ? files : new FileStatus[0];<a name="line.631"></a> -<span class="sourceLineNo">632</span> } catch (FileNotFoundException e) {<a name="line.632"></a> -<span class="sourceLineNo">633</span> return new FileStatus[0];<a name="line.633"></a> -<span class="sourceLineNo">634</span> }<a name="line.634"></a> -<span class="sourceLineNo">635</span> }<a name="line.635"></a> -<span class="sourceLineNo">636</span><a name="line.636"></a> -<span class="sourceLineNo">637</span> private static long getMaxSequenceId(FileStatus[] files) {<a name="line.637"></a> -<span class="sourceLineNo">638</span> long maxSeqId = -1L;<a name="line.638"></a> -<span class="sourceLineNo">639</span> for (FileStatus file : files) {<a name="line.639"></a> -<span class="sourceLineNo">640</span> String fileName = file.getPath().getName();<a name="line.640"></a> -<span class="sourceLineNo">641</span> try {<a name="line.641"></a> -<span class="sourceLineNo">642</span> maxSeqId = Math.max(maxSeqId, Long<a name="line.642"></a> -<span class="sourceLineNo">643</span> .parseLong(fileName.substring(0, fileName.length() - SEQUENCE_ID_FILE_SUFFIX_LENGTH)));<a name="line.643"></a> -<span class="sourceLineNo">644</span> } catch (NumberFormatException ex) {<a name="line.644"></a> -<span class="sourceLineNo">645</span> LOG.warn("Invalid SeqId File Name={}", fileName);<a name="line.645"></a> -<span class="sourceLineNo">646</span> }<a name="line.646"></a> -<span class="sourceLineNo">647</span> }<a name="line.647"></a> -<span class="sourceLineNo">648</span> return maxSeqId;<a name="line.648"></a> -<span class="sourceLineNo">649</span> }<a name="line.649"></a> -<span class="sourceLineNo">650</span><a name="line.650"></a> -<span class="sourceLineNo">651</span> /**<a name="line.651"></a> -<span class="sourceLineNo">652</span> * Get the max sequence id which is stored in the region directory. -1 if none.<a name="line.652"></a> -<span class="sourceLineNo">653</span> */<a name="line.653"></a> -<span class="sourceLineNo">654</span> public static long getMaxRegionSequenceId(FileSystem fs, Path regionDir) throws IOException {<a name="line.654"></a> -<span class="sourceLineNo">655</span> return getMaxSequenceId(getSequenceIdFiles(fs, regionDir));<a name="line.655"></a> -<span class="sourceLineNo">656</span> }<a name="line.656"></a> -<span class="sourceLineNo">657</span><a name="line.657"></a> -<span class="sourceLineNo">658</span> /**<a name="line.658"></a> -<span class="sourceLineNo">659</span> * Create a file with name as region's max sequence id<a name="line.659"></a> -<span class="sourceLineNo">660</span> */<a name="line.660"></a> -<span class="sourceLineNo">661</span> public static void writeRegionSequenceIdFile(FileSystem fs, Path regionDir, long newMaxSeqId)<a name="line.661"></a> -<span class="sourceLineNo">662</span> throws IOException {<a name="line.662"></a> -<span class="sourceLineNo">663</span> FileStatus[] files = getSequenceIdFiles(fs, regionDir);<a name="line.663"></a> -<span class="sourceLineNo">664</span> long maxSeqId = getMaxSequenceId(files);<a name="line.664"></a> -<span class="sourceLineNo">665</span> if (maxSeqId > newMaxSeqId) {<a name="line.665"></a> -<span class="sourceLineNo">666</span> throw new IOException("The new max sequence id " + newMaxSeqId +<a name="line.666"></a> -<span class="sourceLineNo">667</span> " is less than the old max sequence id " + maxSeqId);<a name="line.667"></a> -<span class="sourceLineNo">668</span> }<a name="line.668"></a> -<span class="sourceLineNo">669</span> // write a new seqId file<a name="line.669"></a> -<span class="sourceLineNo">670</span> Path newSeqIdFile = new Path(WALSplitter.getRegionDirRecoveredEditsDir(regionDir),<a name="line.670"></a> -<span class="sourceLineNo">671</span> newMaxSeqId + SEQUENCE_ID_FILE_SUFFIX);<a name="line.671"></a> -<span class="sourceLineNo">672</span> if (newMaxSeqId != maxSeqId) {<a name="line.672"></a> -<span class="sourceLineNo">673</span> try {<a name="line.673"></a> -<span class="sourceLineNo">674</span> if (!fs.createNewFile(newSeqIdFile) && !fs.exists(newSeqIdFile)) {<a name="line.674"></a> -<span class="sourceLineNo">675</span> throw new IOException("Failed to create SeqId file:" + newSeqIdFile);<a name="line.675"></a> -<span class="sourceLineNo">676</span> }<a name="line.676"></a> -<span class="sourceLineNo">677</span> LOG.debug("Wrote file={}, newMaxSeqId={}, maxSeqId={}", newSeqIdFile, newMaxSeqId,<a name="line.677"></a> -<span class="sourceLineNo">678</span> maxSeqId);<a name="line.678"></a> -<span class="sourceLineNo">679</span> } catch (FileAlreadyExistsException ignored) {<a name="line.679"></a> -<span class="sourceLineNo">680</span> // latest hdfs throws this exception. it's all right if newSeqIdFile already exists<a name="line.680"></a> -<span class="sourceLineNo">681</span> }<a name="line.681"></a> -<span class="sourceLineNo">682</span> }<a name="line.682"></a> -<span class="sourceLineNo">683</span> // remove old ones<a name="line.683"></a> -<span class="sourceLineNo">684</span> for (FileStatus status : files) {<a name="line.684"></a> -<span class="sourceLineNo">685</span> if (!newSeqIdFile.equals(status.getPath())) {<a name="line.685"></a> -<span class="sourceLineNo">686</span> fs.delete(status.getPath(), false);<a name="line.686"></a> -<span class="sourceLineNo">687</span> }<a name="line.687"></a> -<span class="sourceLineNo">688</span> }<a name="line.688"></a> -<span class="sourceLineNo">689</span> }<a name="line.689"></a> -<span class="sourceLineNo">690</span><a name="line.690"></a> -<span class="sourceLineNo">691</span> /**<a name="line.691"></a> -<span class="sourceLineNo">692</span> * Create a new {@link Reader} for reading logs to split.<a name="line.692"></a> -<span class="sourceLineNo">693</span> *<a name="line.693"></a> -<span class="sourceLineNo">694</span> * @param file<a name="line.694"></a> -<span class="sourceLineNo">695</span> * @return A new Reader instance, caller should close<a name="line.695"></a> -<span class="sourceLineNo">696</span> * @throws IOException<a name="line.696"></a> -<span class="sourceLineNo">697</span> * @throws CorruptedLogFileException<a name="line.697"></a> -<span class="sourceLineNo">698</span> */<a name="line.698"></a> -<span class="sourceLineNo">699</span> protected Reader getReader(FileStatus file, boolean skipErrors, CancelableProgressable reporter)<a name="line.699"></a> -<span class="sourceLineNo">700</span> throws IOException, CorruptedLogFileException {<a name="line.700"></a> -<span class="sourceLineNo">701</span> Path path = file.getPath();<a name="line.701"></a> -<span class="sourceLineNo">702</span> long length = file.getLen();<a name="line.702"></a> -<span class="sourceLineNo">703</span> Reader in;<a name="line.703"></a> -<span class="sourceLineNo">704</span><a name="line.704"></a> -<span class="sourceLineNo">705</span> // Check for possibly empty file. With appends, currently Hadoop reports a<a name="line.705"></a> -<span class="sourceLineNo">706</span> // zero length even if the file has been sync'd. Revisit if HDFS-376 or<a name="line.706"></a> -<span class="sourceLineNo">707</span> // HDFS-878 is committed.<a name="line.707"></a> -<span class="sourceLineNo">708</span> if (length <= 0) {<a name="line.708"></a> -<span class="sourceLineNo">709</span> LOG.warn("File {} might be still open, length is 0", path);<a name="line.709"></a> -<span class="sourceLineNo">710</span> }<a name="line.710"></a> -<span class="sourceLineNo">711</span><a name="line.711"></a> -<span class="sourceLineNo">712</span> try {<a name="line.712"></a> -<span class="sourceLineNo">713</span> FSUtils.getInstance(fs, conf).recoverFileLease(fs, path, conf, reporter);<a name="line.713"></a> -<span class="sourceLineNo">714</span> try {<a name="line.714"></a> -<span class="sourceLineNo">715</span> in = getReader(path, reporter);<a name="line.715"></a> -<span class="sourceLineNo">716</span> } catch (EOFException e) {<a name="line.716"></a> -<span class="sourceLineNo">717</span> if (length <= 0) {<a name="line.717"></a> -<span class="sourceLineNo">718</span> // TODO should we ignore an empty, not-last log file if skip.errors<a name="line.718"></a> -<span class="sourceLineNo">719</span> // is false? Either way, the caller should decide what to do. E.g.<a name="line.719"></a> -<span class="sourceLineNo">720</span> // ignore if this is the last log in sequence.<a name="line.720"></a> -<span class="sourceLineNo">721</span> // TODO is this scenario still possible if the log has been<a name="line.721"></a> -<span class="sourceLineNo">722</span> // recovered (i.e. closed)<a name="line.722"></a> -<span class="sourceLineNo">723</span> LOG.warn("Could not open {} for reading. File is empty", path, e);<a name="line.723"></a> -<span class="sourceLineNo">724</span> }<a name="line.724"></a> -<span class="sourceLineNo">725</span> // EOFException being ignored<a name="line.725"></a> -<span class="sourceLineNo">726</span> return null;<a name="line.726"></a> -<span class="sourceLineNo">727</span> }<a name="line.727"></a> -<span class="sourceLineNo">728</span> } catch (IOException e) {<a name="line.728"></a> -<span class="sourceLineNo">729</span> if (e instanceof FileNotFoundException) {<a name="line.729"></a> -<span class="sourceLineNo">730</span> // A wal file may not exist anymore. Nothing can be recovered so move on<a name="line.730"></a> -<span class="sourceLineNo">731</span> LOG.warn("File {} does not exist anymore", path, e);<a name="line.731"></a> -<span class="sourceLineNo">732</span> return null;<a name="line.732"></a> -<span class="sourceLineNo">733</span> }<a name="line.733"></a> -<span class="sourceLineNo">734</span> if (!skipErrors || e instanceof InterruptedIOException) {<a name="line.734"></a> -<span class="sourceLineNo">735</span> throw e; // Don't mark the file corrupted if interrupted, or not skipErrors<a name="line.735"></a> -<span class="sourceLineNo">736</span> }<a name="line.736"></a> -<span class="sourceLineNo">737</span> CorruptedLogFileException t =<a name="line.737"></a> -<span class="sourceLineNo">738</span> new CorruptedLogFileException("skipErrors=true Could not open wal " +<a name="line.738"></a> -<span class="sourceLineNo">739</span> path + " ignoring");<a name="line.739"></a> -<span class="sourceLineNo">740</span> t.initCause(e);<a name="line.740"></a> -<span class="sourceLineNo">741</span> throw t;<a name="line.741"></a> -<span class="sourceLineNo">742</span> }<a name="line.742"></a> -<span class="sourceLineNo">743</span> return in;<a name="line.743"></a> -<span class="sourceLineNo">744</span> }<a name="line.744"></a> -<span class="sourceLineNo">745</span><a name="line.745"></a> -<span class="sourceLineNo">746</span> static private Entry getNextLogLine(Reader in, Path path, boolean skipErrors)<a name="line.746"></a> -<span class="sourceLineNo">747</span> throws CorruptedLogFileException, IOException {<a name="line.747"></a> -<span class="sourceLineNo">748</span> try {<a name="line.748"></a> -<span class="sourceLineNo">749</span> return in.next();<a name="line.749"></a> -<span class="sourceLineNo">750</span> } catch (EOFException eof) {<a name="line.750"></a> -<span class="sourceLineNo">751</span> // truncated files are expected if a RS crashes (see HBASE-2643)<a name="line.751"></a> -<span class="sourceLineNo">752</span> LOG.info("EOF from wal {}. Continuing.", path);<a name="line.752"></a> -<span class="sourceLineNo">753</span> return null;<a name="line.753"></a> -<span class="sourceLineNo">754</span> } catch (IOException e) {<a name="line.754"></a> -<span class="sourceLineNo">755</span> // If the IOE resulted from bad file format,<a name="line.755"></a> -<span class="sourceLineNo">756</span> // then this problem is idempotent and retrying won't help<a name="line.756"></a> -<span class="sourceLineNo">757</span> if (e.getCause() != null &&<a name="line.757"></a> -<span class="sourceLineNo">758</span> (e.getCause() instanceof ParseException ||<a name="line.758"></a> -<span class="sourceLineNo">759</span> e.getCause() instanceof org.apache.hadoop.fs.ChecksumException)) {<a name="line.759"></a> -<span class="sourceLineNo">760</span> LOG.warn("Parse exception from wal {}. Continuing", path, e);<a name="line.760"></a> -<span class="sourceLineNo">761</span> return null;<a name="line.761"></a> -<span class="sourceLineNo">762</span> }<a name="line.762"></a> -<span class="sourceLineNo">763</span> if (!skipErrors) {<a name="line.763"></a> -<span class="sourceLineNo">764</span> throw e;<a name="line.764"></a> -<span class="sourceLineNo">765</span> }<a name="line.765"></a> -<span class="sourceLineNo">766</span> CorruptedLogFileException t =<a name="line.766"></a> -<span class="sourceLineNo">767</span> new CorruptedLogFileException("skipErrors=true Ignoring exception" +<a name="line.767"></a> -<span class="sourceLineNo">768</span> " while parsing wal " + path + ". Marking as corrupted");<a name="line.768"></a> -<span class="sourceLineNo">769</span> t.initCause(e);<a name="line.769"></a> -<span class="sourceLineNo">770</span> throw t;<a name="line.770"></a> -<span class="sourceLineNo">771</span> }<a name="line.771"></a> -<span class="sourceLineNo">772</span> }<a name="line.772"></a> -<span class="sourceLineNo">773</span><a name="line.773"></a> -<span class="sourceLineNo">774</span> /**<a name="line.774"></a> -<span class="sourceLineNo">775</span> * Create a new {@link Writer} for writing log splits.<a name="line.775"></a> -<span class="sourceLineNo">776</span> * @return a new Writer instance, caller should close<a name="line.776"></a> -<span class="sourceLineNo">777</span> */<a name="line.777"></a> -<span class="sourceLineNo">778</span> protected Writer createWriter(Path logfile)<a name="line.778"></a> -<span class="sourceLineNo">779</span> throws IOException {<a name="line.779"></a> -<span class="sourceLineNo">780</span> return walFactory.createRecoveredEditsWriter(fs, logfile);<a name="line.780"></a> -<span class="sourceLineNo">781</span> }<a name="line.781"></a> -<span class="sourceLineNo">782</span><a name="line.782"></a> -<span class="sourceLineNo">783</span> /**<a name="line.783"></a> -<span class="sourceLineNo">784</span> * Create a new {@link Reader} for reading logs to split.<a name="line.784"></a> -<span class="sourceLineNo">785</span> * @return new Reader instance, caller should close<a name="line.785"></a> -<span class="sourceLineNo">786</span> */<a name="line.786"></a> -<span class="sourceLineNo">787</span> protected Reader getReader(Path curLogFile, CancelableProgressable reporter) throws IOException {<a name="line.787"></a> -<span class="sourceLineNo">788</span> return walFactory.createReader(fs, curLogFile, reporter);<a name="line.788"></a> -<span class="sourceLineNo">789</span> }<a name="line.789"></a> -<span class="sourceLineNo">790</span><a name="line.790"></a> -<span class="sourceLineNo">791</span> /**<a name="line.791"></a> -<span class="sourceLineNo">792</span> * Get current open writers<a name="line.792"></a> -<span class="sourceLineNo">793</span> */<a name="line.793"></a> -<span class="sourceLineNo">794</span> private int getNumOpenWriters() {<a name="line.794"></a> -<span class="sourceLineNo">795</span> int result = 0;<a name="line.795"></a> -<span class="sourceLineNo">796</span> if (this.outputSink != null) {<a name="line.796"></a> -<span class="sourceLineNo">797</span> result += this.outputSink.getNumOpenWriters();<a name="line.797"></a> -<span class="sourceLineNo">798</span> }<a name="line.798"></a> -<span class="sourceLineNo">799</span> return result;<a name="line.799"></a> -<span class="sourceLineNo">800</span> }<a name="line.800"></a> -<span class="sourceLineNo">801</span><a name="line.801"></a> -<span class="sourceLineNo">802</span> /**<a name="line.802"></a> -<span class="sourceLineNo">803</span> * Contains some methods to control WAL-entries producer / consumer interactions<a name="line.803"></a> -<span class="sourceLineNo">804</span> */<a name="line.804"></a> -<span class="sourceLineNo">805</span> public static class PipelineController {<a name="line.805"></a> -<span class="sourceLineNo">806</span> // If an exception is thrown by one of the other threads, it will be<a name="line.806"></a> -<span class="sourceLineNo">807</span> // stored here.<a name="line.807"></a> -<span class="sourceLineNo">808</span> AtomicReference<Throwable> thrown = new AtomicReference<>();<a name="line.808"></a> -<span class="sourceLineNo">809</span><a name="line.809"></a> -<span class="sourceLineNo">810</span> // Wait/notify for when data has been produced by the writer thread,<a name="line.810"></a> -<span class="sourceLineNo">811</span> // consumed by the reader thread, or an exception occurred<a name="line.811"></a> -<span class="sourceLineNo">812</span> public final Object dataAvailable = new Object();<a name="line.812"></a> -<span class="sourceLineNo">813</span><a name="line.813"></a> -<span class="sourceLineNo">814</span> void writerThreadError(Throwable t) {<a name="line.814"></a> -<span class="sourceLineNo">815</span> thrown.compareAndSet(null, t);<a name="line.815"></a> -<span class="sourceLineNo">816</span> }<a name="line.816"></a> -<span class="sourceLineNo">817</span><a name="line.817"></a> -<span class="sourceLineNo">818</span> /**<a name="line.818"></a> -<span class="sourceLineNo">819</span> * Check for errors in the writer threads. If any is found, rethrow it.<a name="line.819"></a> -<span class="sourceLineNo">820</span> */<a name="line.820"></a> -<span class="sourceLineNo">821</span> void checkForErrors() throws IOException {<a name="line.821"></a> -<span class="sourceLineNo">822</span> Throwable thrown = this.thrown.get();<a name="line.822"></a> -<span class="sourceLineNo">823</span> if (thrown == null) return;<a name="line.823"></a> -<span class="sourceLineNo">824</span> if (thrown instanceof IOException) {<a name="line.824"></a> -<span class="sourceLineNo">825</span> throw new IOException(thrown);<a name="line.825"></a> -<span class="sourceLineNo">826</span> } else {<a name="line.826"></a> -<span class="sourceLineNo">827</span> throw new RuntimeException(thrown);<a name="line.827"></a> -<span class="sourceLineNo">828</span> }<a name="line.828"></a> -<span class="sourceLineNo">829</span> }<a name="line.829"></a> -<span class="sourceLineNo">830</span> }<a name="line.830"></a> -<span class="sourceLineNo">831</span><a name="line.831"></a> -<span class="sourceLineNo">832</span> /**<a name="line.832"></a> -<span class="sourceLineNo">833</span> * Class which accumulates edits and separates them into a buffer per region<a name="line.833"></a> -<span class="sourceLineNo">834</span> * while simultaneously accounting RAM usage. Blocks if the RAM usage crosses<a name="line.834"></a> -<span class="sourceLineNo">835</span> * a predefined threshold.<a name="line.835"></a> -<span class="sourceLineNo">836</span> *<a name="line.836"></a> -<span class="sourceLineNo">837</span> * Writer threads then pull region-specific buffers from this class.<a name="line.837"></a> -<span class="sourceLineNo">838</span> */<a name="line.838"></a> -<span class="sourceLineNo">839</span> public static class EntryBuffers {<a name="line.839"></a> -<span class="sourceLineNo">840</span> PipelineController controller;<a name="line.840"></a> -<span class="sourceLineNo">841</span><a name="line.841"></a> -<span class="sourceLineNo">842</span> Map<byte[], RegionEntryBuffer> buffers = new TreeMap<>(Bytes.BYTES_COMPARATOR);<a name="line.842"></a> -<span class="sourceLineNo">843</span><a name="line.843"></a> -<span class="sourceLineNo">844</span> /* Track which regions are currently in the middle of writing. We don't allow<a name="line.844"></a> -<span class="sourceLineNo">845</span> an IO thread to pick up bytes from a region if we're already writing<a name="line.845"></a> -<span class="sourceLineNo">846</span> data for that region in a different IO thread. */<a name="line.846"></a> -<span class="sourceLineNo">847</span> Set<byte[]> currentlyWriting = new TreeSet<>(Bytes.BYTES_COMPARATOR);<a name="line.847"></a> -<span class="sourceLineNo">848</span><a name="line.848"></a> -<span class="sourceLineNo">849</span> long totalBuffered = 0;<a name="line.849"></a> -<span class="sourceLineNo">850</span> long maxHeapUsage;<a name="line.850"></a> -<span class="sourceLineNo">851</span> boolean splitWriterCreationBounded;<a name="line.851"></a> -<span class="sourceLineNo">852</span><a name="line.852"></a> -<span class="sourceLineNo">853</span> public EntryBuffers(PipelineController controller, long maxHeapUsage) {<a name="line.853"></a> -<span class="sourceLineNo">854</span> this(controller, maxHeapUsage, false);<a name="line.854"></a> -<span class="sourceLineNo">855</span> }<a name="line.855"></a> -<span class="sourceLineNo">856</span><a name="line.856"></a> -<span class="sourceLineNo">857</span> public EntryBuffers(PipelineController controller, long maxHeapUsage,<a name="line.857"></a> -<span class="sourceLineNo">858</span> boolean splitWriterCreationBounded){<a name="line.858"></a> -<span class="sourceLineNo">859</span> this.controller = controller;<a name="line.859"></a> -<span class="sourceLineNo">860</span> this.maxHeapUsage = maxHeapUsage;<a name="line.860"></a> -<span class="sourceLineNo">861</span> this.splitWriterCreationBounded = splitWriterCreationBounded;<a name="line.861"></a> -<span class="sourceLineNo">862</span> }<a name="line.862"></a> -<span class="sourceLineNo">863</span><a name="line.863"></a> -<span class="sourceLineNo">864</span> /**<a name="line.864"></a> -<span class="sourceLineNo">865</span> * Append a log entry into the corresponding region buffer.<a name="line.865"></a> -<span class="sourceLineNo">866</span> * Blocks if the total heap usage has crossed the specified threshold.<a name="line.866"></a> -<span class="sourceLineNo">867</span> *<a name="line.867"></a> -<span class="sourceLineNo">868</span> * @throws InterruptedException<a name="line.868"></a> -<span class="sourceLineNo">869</span> * @throws IOException<a name="line.869"></a> -<span class="sourceLineNo">870</span> */<a name="line.870"></a> -<span class="sourceLineNo">871</span> public void appendEntry(Entry entry) throws InterruptedException, IOException {<a name="line.871"></a> -<span class="sourceLineNo">872</span> WALKey key = entry.getKey();<a name="line.872"></a> -<span class="sourceLineNo">873</span><a name="line.873"></a> -<span class="sourceLineNo">874</span> RegionEntryBuffer buffer;<a name="line.874"></a> -<span class="sourceLineNo">875</span> long incrHeap;<a name="line.875"></a> -<span class="sourceLineNo">876</span> synchronized (this) {<a name="line.876"></a> -<span class="sourceLineNo">877</span> buffer = buffers.get(key.getEncodedRegionName());<a name="line.877"></a> -<span class="sourceLineNo">878</span> if (buffer == null) {<a name="line.878"></a> -<span class="sourceLineNo">879</span> buffer = new RegionEntryBuffer(key.getTableName(), key.getEncodedRegionName());<a name="line.879"></a> -<span class="sourceLineNo">880</span> buffers.put(key.getEncodedRegionName(), buffer);<a name="line.880"></a> -<span class="sourceLineNo">881</span> }<a name="line.881"></a> -<span class="sourceLineNo">882</span> incrHeap= buffer.appendEntry(entry);<a name="line.882"></a> -<span class="sourceLineNo">883</span> }<a name="line.883"></a> -<span class="sourceLineNo">884</span><a name="line.884"></a> -<span class="sourceLineNo">885</span> // If we crossed the chunk threshold, wait for more space to be available<a name="line.885"></a> -<span class="sourceLineNo">886</span> synchronized (controller.dataAvailable) {<a name="line.886"></a> -<span class="sourceLineNo">887</span> totalBuffered += incrHeap;<a name="line.887"></a> -<span class="sourceLineNo">888</span> while (totalBuffered > maxHeapUsage && controller.thrown.get() == null) {<a name="line.888"></a> -<span class="sourceLineNo">889</span> LOG.debug("Used {} bytes of buffered edits, waiting for IO threads", totalBuffered);<a name="line.889"></a> -<span class="sourceLineNo">890</span> controller.dataAvailable.wait(2000);<a name="line.890"></a> -<span class="sourceLineNo">891</span> }<a name="line.891"></a> -<span class="sourceLineNo">892</span> controller.dataAvailable.notifyAll();<a name="line.892"></a> -<span class="sourceLineNo">893</span> }<a name="line.893"></a> -<span class="sourceLineNo">894</span> controller.checkForErrors();<a name="line.894"></a> -<span class="sourceLineNo">895</span> }<a name="line.895"></a> -<span class="sourceLineNo">896</span><a name="line.896"></a> -<span class="sourceLineNo">897</span> /**<a name="line.897"></a> -<span class="sourceLineNo">898</span> * @return RegionEntryBuffer a buffer of edits to be written.<a name="line.898"></a> -<span class="sourceLineNo">899</span> */<a name="line.899"></a> -<span class="sourceLineNo">900</span> synchronized RegionEntryBuffer getChunkToWrite() {<a name="line.900"></a> -<span class="sourceLineNo">901</span> // The core part of limiting opening writers is it doesn't return chunk only if the<a name="line.901"></a> -<span class="sourceLineNo">902</span> // heap size is over maxHeapUsage. Thus it doesn't need to create a writer for each<a name="line.902"></a> -<span class="sourceLineNo">903</span> // region during splitting. It will flush all the logs in the buffer after splitting<a name="line.903"></a> -<span class="sourceLineNo">904</span> // through a threadpool, which means the number of writers it created is under control.<a name="line.904"></a> -<span class="sourceLineNo">905</span> if (splitWriterCreationBounded && totalBuffered < maxHeapUsage) {<a name="line.905"></a> -<span class="sourceLineNo">906</span> return null;<a name="line.906"></a> -<span class="sourceLineNo">907</span> }<a name="line.907"></a> -<span class="sourceLineNo">908</span> long biggestSize = 0;<a name="line.908"></a> -<span class="sourceLineNo">909</span> byte[] biggestBufferKey = null;<a name="line.909"></a> -<span class="sourceLineNo">910</span><a name="line.910"></a> -<span class="sourceLineNo">911</span> for (Map.Entry<byte[], RegionEntryBuffer> entry : buffers.entrySet()) {<a name="line.911"></a> -<span class="sourceLineNo">912</span> long size = entry.getValue().heapSize();<a name="line.912"></a> -<span class="sourceLineNo">913</span> if (size > biggestSize && (!currentlyWriting.contains(entry.getKey()))) {<a name="line.913"></a> -<span class="sourceLineNo">914</span> biggestSize = size;<a name="line.914"></a> -<span class="sourceLineNo">915</span> biggestBufferKey = entry.getKey();<a name="line.915"></a> -<span class="sourceLineNo">916</span> }<a name="line.916"></a> -<span class="sourceLineNo">917</span> }<a name="line.917"></a> -<span class="sourceLineNo">918</span> if (biggestBufferKey == null) {<a name="line.918"></a> -<span class="sourceLineNo">919</span> return null;<a name="line.919"></a> -<span class="sourceLineNo">920</span> }<a name="line.920"></a> -<span class="sourceLineNo">921</span><a name="line.921"></a> -<span class="sourceLineNo">922</span> RegionEntryBuffer buffer = buffers.remove(biggestBufferKey);<a name="line.922"></a> -<span class="sourceLineNo">923</span> currentlyWriting.add(biggestBufferKey);<a name="line.923"></a> -<span class="sourceLineNo">924</span> return buffer;<a name="line.924"></a> -<span class="sourceLineNo">925</span> }<a name="line.925"></a> -<span class="sourceLineNo">926</span><a name="line.926"></a> -<span class="sourceLineNo">927</span> void doneWriting(RegionEntryBuffer buffer) {<a name="line.927"></a> -<span class="sourceLineNo">928</span> synchronized (this) {<a name="line.928"></a> -<span class="sourceLineNo">929</span> boolean removed = currentlyWriting.remove(buffer.encodedRegionName);<a name="line.929"></a> -<span class="sourceLineNo">930</span> assert removed;<a name="line.930"></a> -<span class="sourceLineNo">931</span> }<a name="line.931"></a> -<span class="sourceLineNo">932</span> long size = buffer.heapSize();<a name="line.932"></a> -<span class="sourceLineNo">933</span><a name="line.933"></a> -<span class="sourceLineNo">934</span> synchronized (controller.dataAvailable) {<a name="line.934"></a> -<span class="sourceLineNo">935</span> totalBuffered -= size;<a name="line.935"></a> -<span class="sourceLineNo">936</span> // We may unblock writers<a name="line.936"></a> -<span class="sourceLineNo">937</span> controller.dataAvailable.notifyAll();<a name="line.937"></a> -<span class="sourceLineNo">938</span> }<a name="line.938"></a> -<span class="sourceLineNo">939</span> }<a name="line.939"></a> -<span class="sourceLineNo">940</span><a name="line.940"></a> -<span class="sourceLineNo">941</span> synchronized boolean isRegionCurrentlyWriting(byte[] region) {<a name="line.941"></a> -<span class="sourceLineNo">942</span> return currentlyWriting.contains(region);<a name="line.942"></a> -<span class="sourceLineNo">943</span> }<a name="line.943"></a> -<span class="sourceLineNo">944</span><a name="line.944"></a> -<span class="sourceLineNo">945</span> public void waitUntilDrained() {<a name="line.945"></a> -<span class="sourceLineNo">946</span> synchronized (controller.dataAvailable) {<a name="line.946"></a> -<span class="sourceLineNo">947</span> while (totalBuffered > 0) {<a name="line.947"></a> -<span class="sourceLineNo">948</span> try {<a name="line.948"></a> -<span class="sourceLineNo">949</span> controller.dataAvailable.wait(2000);<a name="line.949"></a> -<span class="sourceLineNo">950</span> } catch (InterruptedException e) {<a name="line.950"></a> -<span class="sourceLineNo">951</span> LOG.warn("Got interrupted while waiting for EntryBuffers is drained");<a name="line.951"></a> -<span class="sourceLineNo">952</span> Thread.interrupted();<a name="line.952"></a> -<span class="sourceLineNo">953</span> break;<a name="line.953"></a> -<span class="sourceLineNo">954</span> }<a name="line.954"></a> -<span class="sourceLineNo">955</span> }<a name="line.955"></a> -<span class="sourceLineNo">956</span> }<a name="line.956"></a> -<span class="sourceLineNo">957</span> }<a name="line.957"></a> -<span class="sourceLineNo">958</span> }<a name="line.958"></a> -<span class="sourceLineNo">959</span><a name="line.959"></a> -<span class="sourceLineNo">960</span> /**<a name="line.960"></a> -<span class="sourceLineNo">961</span> * A buffer of some number of edits for a given region.<a name="line.961"></a> -<span class="sourceLineNo">962</span> * This accumulates edits and also provides a memory optimization in order to<a name="line.962"></a> -<span class="sourceLineNo">963</span> * share a single byte array instance for the table and region name.<a name="line.963"></a> -<span class="sourceLineNo">964</span> * Also tracks memory usage of the accumulated edits.<a name="line.964"></a> -<span class="sourceLineNo">965</span> */<a name="line.965"></a> -<span class="sourceLineNo">966</span> public static class RegionEntryBuffer implements HeapSize {<a name="line.966"></a> -<span class="sourceLineNo">967</span> long heapInBuffer = 0;<a name="line.967"></a> -<span class="sourceLineNo">968</span> List<Entry> entryBuffer;<a name="line.968"></a> -<span class="sourceLineNo">969</span> TableName tableName;<a name="line.969"></a> -<span class="sourceLineNo">970</span> byte[] encodedRegionName;<a name="line.970"></a> -<span class="sourceLineNo">971</span><a name="line.971"></a> -<span class="sourceLineNo">972</span> RegionEntryBuffer(TableName tableName, byte[] region) {<a name="line.972"></a> -<span class="sourceLineNo">973</span> this.tableName = tableName;<a name="line.973"></a> -<span class="sourceLineNo">974</span> this.encodedRegionName = region;<a name="line.974"></a> -<span class="sourceLineNo">975</span> this.entryBuffer = new ArrayList<>();<a name="line.975"></a> -<span class="sourceLineNo">976</span> }<a name="line.976"></a> -<span class="sourceLineNo">977</span><a name="line.977"></a> -<span class="sourceLineNo">978</span> long appendEntry(Entry entry) {<a name="line.978"></a> -<span class="sourceLineNo">979</span> internify(entry);<a name="line.979"></a> -<span class="sourceLineNo">980</span> entryBuffer.add(entry);<a name="line.980"></a> -<span class="sourceLineNo">981</span> long incrHeap = entry.getEdit().heapSize() +<a name="line.981"></a> -<span class="sourceLineNo">982</span> ClassSize.align(2 * ClassSize.REFERENCE) + // WALKey pointers<a name="line.982"></a> -<span class="sourceLineNo">983</span> 0; // TODO linkedlist entry<a name="line.983"></a> -<span class="sourceLineNo">984</span> heapInBuffer += incrHeap;<a name="line.984"></a> -<span class="sourceLineNo">985</span> return incrHeap;<a name="line.985"></a> -<span class="sourceLineNo">986</span> }<a name="line.986"></a> -<span class="sourceLineNo">987</span><a name="line.987"></a> -<span class="sourceLineNo">988</span> private void internify(Entry entry) {<a name="line.988"></a> -<span class="sourceLineNo">989</span> WALKeyImpl k = entry.getKey();<a name="line.989"></a> -<span class="sourceLineNo">990</span> k.internTableName(this.tableName);<a name="line.990"></a> -<span class="sourceLineNo">991</span> k.internEncodedRegionName(this.encodedRegionName);<a name="line.991"></a> -<span class="sourceLineNo">992</span> }<a name="line.992"></a> -<span class="sourceLineNo">993</span><a name="line.993"></a> -<span class="sourceLineNo">994</span> @Override<a name="line.994"></a> -<span class="sourceLineNo">995</span> public long heapSize() {<a name="line.995"></a> -<span class="sourceLineNo">996</span> return heapInBuffer;<a name="line.996"></a> -<span class="sourceLineNo">997</span> }<a name="line.997"></a> -<span class="sourceLineNo">998</span><a name="line.998"></a> -<span class="sourceLineNo">999</span> public byte[] getEncodedRegionName() {<a name="line.999"></a> -<span class="sourceLineNo">1000</span> return encodedRegionName;<a name="line.1000"></a> -<span class="sourceLineNo">1001</span> }<a name="line.1001"></a> -<span class="sourceLineNo">1002</span><a name="line.1002"></a> -<span class="sourceLineNo">1003</span> public List<Entry> getEntryBuffer() {<a name="line.1003"></a> -<span class="sourceLineNo">1004</span> return entryBuffer;<a name="line.1004"></a> -<span class="sourceLineNo">1005</span> }<a name="line.1005"></a> -<span class="sourceLineNo">1006</span><a name="line.1006"></a> -<span class="sourceLineNo">1007</span> public TableName getTableName() {<a name="line.1007"></a> -<span class="sourceLineNo">1008</span> return tableName;<a name="line.1008"></a> -<span class="sourceLineNo">1009</span> }<a name="line.1009"></a> -<span class="sourceLineNo">1010</span> }<a name="line.1010"></a> -<span class="sourceLineNo">1011</span><a name="line.1011"></a> -<span class="sourceLineNo">1012</span> public static class WriterThread extends Thread {<a name="line.1012"></a> -<span class="sourceLineNo">1013</span> private volatile boolean shouldStop = false;<a name="line.1013"></a> -<span class="sourceLineNo">1014</span> private PipelineController controller;<a name="line.1014"></a> -<span class="sourceLineNo">1015</span> private EntryBuffers entryBuffers;<a name="line.1015"></a> -<span class="sourceLineNo">1016</span> private OutputSink outputSink = null;<a name="line.1016"></a> -<span class="sourceLineNo">1017</span><a name="line.1017"></a> -<span class="sourceLineNo">1018</span> WriterThread(PipelineController controller, EntryBuffers entryBuffers, OutputSink sink, int i){<a name="line.1018"></a> -<span class="sourceLineNo">1019</span> super(Thread.currentThread().getName() + "-Writer-" + i);<a name="line.1019"></a> -<span class="sourceLineNo">1020</span> this.controller = controller;<a name="line.1020"></a> -<span class="sourceLineNo">1021</span> this.entryBuffers = entryBuffers;<a name="line.1021"></a> -<span class="sourceLineNo">1022</span> outputSink = sink;<a name="line.1022"></a> -<span class="sourceLineNo">1023</span> }<a name="line.1023"></a> -<span class="sourceLineNo">1024</span><a name="line.1024"></a> -<span class="sourceLineNo">1025</span> @Override<a name="line.1025"></a> -<span class="sourceLineNo">1026</span> public void run() {<a name="line.1026"></a> -<span class="sourceLineNo">1027</span> try {<a name="line.1027"></a> -<span class="sourceLineNo">1028</span> doRun();<a name="line.1028"></a> -<span class="sourceLineNo">1029</span> } catch (Throwable t) {<a name="line.1029"></a> -<span class="sourceLineNo">1030</span> LOG.error("Exiting thread", t);<a name="line.1030"></a> -<span class="sourceLineNo">1031</span> controller.writerThreadError(t);<a name="line.1031"></a> -<span class="sourceLineNo">1032</span> }<a name="line.1032"></a> -<span class="sourceLineNo">1033</span> }<a name="line.1033"></a> -<span class="sourceLineNo">1034</span><a name="line.1034"></a> -<span class="sourceLineNo">1035</span> private void doRun() throws IOException {<a name="line.1035"></a> -<span class="sourceLineNo">1036</span> LOG.trace("Writer thread starting");<a name="line.1036"></a> -<span class="sourceLineNo">1037</span> while (true) {<a name="line.1037"></a> -<span class="sourceLineNo">1038</span> RegionEntryBuffer buffer = entryBuffers.getChunkToWrite();<a name="line.1038"></a> -<span class="sourceLineNo">1039</span> if (buffer == null) {<a name="line.1039"></a> -<span class="sourceLineNo">1040</span> // No data currently available, wait on some more to show up<a name="line.1040"></a> -<span class="sourceLineNo">1041</span> synchronized (controller.dataAvailable) {<a name="line.1041"></a> -<span class="sourceLineNo">1042</span> if (shouldStop && !this.outputSink.flush()) {<a name="line.1042"></a> -<span class="sourceLineNo">1043</span> return;<a name="line.1043"></a> -<span class="sourceLineNo">1044</span> }<a name="line.1044"></a> -<span class="sourceLineNo">1045</span> try {<a name="line.1045"></a> -<span class="sourceLineNo">1046</span> controller.dataAvailable.wait(500);<a name="line.1046"></a> -<span class="sourceLineNo">1047</span> } catch (InterruptedException ie) {<a name="line.1047"></a> -<span class="sourceLineNo">1048</span> if (!shouldStop) {<a name="line.1048"></a> -<span class="sourceLineNo">1049</span> throw new RuntimeException(ie);<a name="line.1049"></a> -<span class="sourceLineNo">1050</span> }<a name="line.1050"></a> -<span class="sourceLineNo">1051</span> }<a name="line.1051"></a> -<span class="sourceLineNo">1052</span> }<a name="line.1052"></a> -<span class="sourceLineNo">1053</span> continue;<a name="line.1053"></a> -<span class="sourceLineNo">1054</span> }<a name="line.1054"></a> -<span class="sourceLineNo">1055</span><a name="line.1055"></a> -<span class="sourceLineNo">1056</span> assert buffer != null;<a name="line.1056"></a> -<span class="sourceLineNo">1057</span> try {<a name="line.1057"></a> -<span class="sourceLineNo">1058</span> writeBuffer(buffer);<a name="line.1058"></a> -<span class="sourceLineNo">1059</span> } finally {<a name="line.1059"></a> -<span class="sourceLineNo">1060</span> entryBuffers.doneWriting(buffer);<a name="line.1060"></a> -<span class="sourceLineNo">1061</span> }<a name="line.1061"></a> -<span class="sourceLineNo">1062</span> }<a name="line.1062"></a> -<span class="sourceLineNo">1063</span> }<a name="line.1063"></a> -<span class="sourceLineNo">1064</span><a name="line.1064"></a> -<span class="sourceLineNo">1065</span> private void writeBuffer(RegionEntryBuffer buffer) throws IOException {<a name="line.1065"></a> -<span class="sourceLineNo">1066</span> outputSink.append(buffer);<a name="line.1066"></a> -<span class="sourceLineNo">1067</span> }<a name="line.1067"></a> -<span class="sourceLineNo">1068</span><a name="line.1068"></a> -<span class="sourceLineNo">1069</span> void finish() {<a name="line.1069"></a> -<span class="sourceLineNo">1070</span> synchronized (controller.dataAvailable) {<a name="line.1070"></a> -<span class="sourceLineNo">1071</span> shouldStop = true;<a name="line.1071"></a> -<span class="sourceLineNo">1072</span> controller.dataAvailable.notifyAll();<a name="line.1072"></a> -<span class="sourceLineNo">1073</span> }<a name="line.1073"></a> -<span class="sourceLineNo">1074</span> }<a name="line.1074"></a> -<span class="sourceLineNo">1075</span> }<a name="line.1075"></a> -<span class="sourceLineNo">1076</span><a name="line.1076"></a> -<span class="sourceLineNo">1077</span> /**<a name="line.1077"></a> -<span class="sourceLineNo">1078</span> * The following class is an abstraction class to provide a common interface to support<a name="line.1078"></a> -<span class="sourceLineNo">1079</span> * different ways of consuming recovered edits.<a name="line.1079"></a> -<span class="sourceLineNo">1080</span> */<a name="line.1080"></a> -<span class="sourceLineNo">1081</span> public static abstract class OutputSink {<a name="line.1081"></a> -<span class="sourceLineNo">1082</span><a name="line.1082"></a> -<span class="sourceLineNo">1083</span> protected PipelineController controller;<a name="line.1083"></a> -<span class="sourceLineNo">1084</span> protected EntryBuffers entryBuffers;<a name="line.1084"></a> -<span class="sourceLineNo">1085</span><a name="line.1085"></a> -<span class="sourceLineNo">1086</span> protected ConcurrentHashMap<String, SinkWriter> writers = new ConcurrentHashMap<>();<a name="line.1086"></a> -<span class="sourceLineNo">1087</span> protected final ConcurrentHashMap<String, Long> regionMaximumEditLogSeqNum =<a name="line.1087"></a> -<span class="sourceLineNo">1088</span> new ConcurrentHashMap<>();<a name="line.1088"></a> -<span class="sourceLineNo">1089</span><a name="line.1089"></a> +<span class="sourceLineNo">460</span> * @param logEntry<a name="line.460"></a> +<span class="sourceLineNo">461</span> * @param fileNameBeingSplit the file being split currently. Used to generate tmp file name.<a name="line.461"></a> +<span class="sourceLineNo">462</span> * @param conf<a name="line.462"></a> +<span class="sourceLineNo">463</span> * @return Path to file into which to dump split log edits.<a name="line.463"></a> +<span class="sourceLineNo">464</span> * @throws IOException<a name="line.464"></a> +<span class="sourceLineNo">465</span> */<a name="line.465"></a> +<span class="sourceLineNo">466</span> @SuppressWarnings("deprecation")<a name="line.466"></a> +<span class="sourceLineNo">467</span> @VisibleForTesting<a name="line.467"></a> +<span class="sourceLineNo">468</span> static Path getRegionSplitEditsPath(final Entry logEntry, String fileNameBeingSplit,<a name="line.468"></a> +<span class="sourceLineNo">469</span> Configuration conf)<a name="line.469"></a> +<span class="sourceLineNo">470</span> throws IOException {<a name="line.470"></a> +<span class="sourceLineNo">471</span> FileSystem fs = FileSystem.get(conf);<a name="line.471"></a> +<span class="sourceLineNo">472</span> Path rootDir = FSUtils.getRootDir(conf);<a name="line.472"></a> +<span class="sourceLineNo">473</span> Path tableDir = FSUtils.getTableDir(rootDir, logEntry.getKey().getTableName());<a name="line.473"></a> +<span class="sourceLineNo">474</span> String encodedRegionName = Bytes.toString(logEntry.getKey().getEncodedRegionName());<a name="line.474"></a> +<span class="sourceLineNo">475</span> Path regiondir = HRegion.getRegionDir(tableDir, encodedRegionName);<a name="line.475"></a> +<span class="sourceLineNo">476</span> Path dir = getRegionDirRecoveredEditsDir(regiondir);<a name="line.476"></a> +<span class="sourceLineNo">477</span><a name="line.477"></a> +<span class="sourceLineNo">478</span> if (!fs.exists(regiondir)) {<a name="line.478"></a> +<span class="sourceLineNo">479</span> LOG.info("This region's directory does not exist: {}."<a name="line.479"></a> +<span class="sourceLineNo">480</span> + "It is very likely that it was already split so it is "<a name="line.480"></a> +<span class="sourceLineNo">481</span> + "safe to discard those edits.", regiondir);<a name="line.481"></a> +<span class="sourceLineNo">482</span> return null;<a name="line.482"></a> +<span class="sourceLineNo">483</span> }<a name="line.483"></a> +<span class="sourceLineNo">484</span> if (fs.exists(dir) && fs.isFile(dir)) {<a name="line.484"></a> +<span class="sourceLineNo">485</span> Path tmp = new Path("/tmp");<a name="line.485"></a> +<span class="sourceLineNo">486</span> if (!fs.exists(tmp)) {<a name="line.486"></a> +<span class="sourceLineNo">487</span> fs.mkdirs(tmp);<a name="line.487"></a> +<span class="sourceLineNo">488</span> }<a name="line.488"></a> +<span class="sourceLineNo">489</span> tmp = new Path(tmp,<a name="line.489"></a> +<span class="sourceLineNo">490</span> HConstants.RECOVERED_EDITS_DIR + "_" + encodedRegionName);<a name="line.490"></a> +<span class="sourceLineNo">491</span> LOG.warn("Found existing old file: {}. It could be some "<a name="line.491"></a> +<span class="sourceLineNo">492</span> + "leftover of an old installation. It should be a folder instead. "<a name="line.492"></a> +<span class="sourceLineNo">493</span> + "So moving it to {}", dir, tmp);<a name="line.493"></a> +<span class="sourceLineNo">494</span> if (!fs.rename(dir, tmp)) {<a name="line.494"></a> +<span class="sourceLineNo">495</span> LOG.warn("Failed to sideline old file {}", dir);<a name="line.495"></a> +<span class="sourceLineNo">496</span> }<a name="line.496"></a> +<span class="sourceLineNo">497</span> }<a name="line.497"></a> +<span class="sourceLineNo">498</span><a name="line.498"></a> +<span class="sourceLineNo">499</span> if (!fs.exists(dir) && !fs.mkdirs(dir)) {<a name="line.499"></a> +<span class="sourceLineNo">500</span> LOG.warn("mkdir failed on {}", dir);<a name="line.500"></a> +<span class="sourceLineNo">501</span> }<a name="line.501"></a> +<span class="sourceLineNo">502</span> // Append fileBeingSplit to prevent name conflict since we may have duplicate wal entries now.<a name="line.502"></a> +<span class="sourceLineNo">503</span> // Append file name ends with RECOVERED_LOG_TMPFILE_SUFFIX to ensure<a name="line.503"></a> +<span class="sourceLineNo">504</span> // region's replayRecoveredEdits will not delete it<a name="line.504"></a> +<span class="sourceLineNo">505</span> String fileName = formatRecoveredEditsFileName(logEntry.getKey().getSequenceId());<a name="line.505"></a> +<span class="sourceLineNo">506</span> fileName = getTmpRecoveredEditsFileName(fileName + "-" + fileNameBeingSplit);<a name="line.506"></a> +<span class="sourceLineNo">507</span> return new Path(dir, fileName);<a name="line.507"></a> +<span class="sourceLineNo">508</span> }<a name="line.508"></a> +<span class="sourceLineNo">509</span><a name="line.509"></a> +<span class="sourceLineNo">510</span> private static String getTmpRecoveredEditsFileName(String fileName) {<a name="line.510"></a> +<span class="sourceLineNo">511</span> return fileName + RECOVERED_LOG_TMPFILE_SUFFIX;<a name="line.511"></a> +<span class="sourceLineNo">512</span> }<a name="line.512"></a> +<span class="sourceLineNo">513</span><a name="line.513"></a> +<span class="sourceLineNo">514</span> /**<a name="line.514"></a> +<span class="sourceLineNo">515</span> * Get the completed recovered edits file path, renaming it to be by last edit<a name="line.515"></a> +<span class="sourceLineNo">516</span> * in the file from its first edit. Then we could use the name to skip<a name="line.516"></a> +<span class="sourceLineNo">517</span> * recovered edits when doing {@link HRegion#replayRecoveredEditsIfAny}.<a name="line.517"></a> +<span class="sourceLineNo">518</span> * @param srcPath<a name="line.518"></a> +<span class="sourceLineNo">519</span> * @param maximumEditLogSeqNum<a name="line.519"></a> +<span class="sourceLineNo">520</span> * @return dstPath take file's last edit log seq num as the name<a name="line.520"></a> +<span class="sourceLineNo">521</span> */<a name="line.521"></a> +<span class="sourceLineNo">522</span> private static Path getCompletedRecoveredEditsFilePath(Path srcPath,<a name="line.522"></a> +<span class="sourceLineNo">523</span> long maximumEditLogSeqNum) {<a name="line.523"></a> +<span class="sourceLineNo">524</span> String fileName = formatRecoveredEditsFileName(maximumEditLogSeqNum);<a name="line.524"></a> +<span class="sourceLineNo">525</span> return new Path(srcPath.getParent(), fileName);<a name="line.525"></a> +<span class="sourceLineNo">526</span> }<a name="line.526"></a> +<span class="sourceLineNo">527</span><a name="line.527"></a> +<span class="sourceLineNo">528</span> @VisibleForTesting<a name="line.528"></a> +<span class="sourceLineNo">529</span> static String formatRecoveredEditsFileName(final long seqid) {<a name="line.529"></a> +<span class="sourceLineNo">530</span> return String.format("%019d", seqid);<a name="line.530"></a> +<span class="sourceLineNo">531</span> }<a name="line.531"></a> +<span class="sourceLineNo">532</span><a name="line.532"></a> +<span class="sourceLineNo">533</span> private static final Pattern EDITFILES_NAME_PATTERN = Pattern.compile("-?[0-9]+");<a name="line.533"></a> +<span class="sourceLineNo">534</span> private static final String RECOVERED_LOG_TMPFILE_SUFFIX = ".temp";<a name="line.534"></a> +<span class="sourceLineNo">535</span><a name="line.535"></a> +<span class="sourceLineNo">536</span> /**<a name="line.536"></a> +<span class="sourceLineNo">537</span> * @param regiondir<a name="line.537"></a> +<span class="sourceLineNo">538</span> * This regions directory in the filesystem.<a name="line.538"></a> +<span class="sourceLineNo">539</span> * @return The directory that holds recovered edits files for the region<a name="line.539"></a> +<span class="sourceLineNo">540</span> * <code>regiondir</code><a name="line.540"></a> +<span class="sourceLineNo">541</span> */<a name="line.541"></a> +<span class="sourceLineNo">542</span> public static Path getRegionDirRecoveredEditsDir(final Path regiondir) {<a name="line.542"></a> +<span class="sourceLineNo">543</span> return new Path(regiondir, HConstants.RECOVERED_EDITS_DIR);<a name="line.543"></a> +<span class="sourceLineNo">544</span> }<a name="line.544"></a> +<span class="sourceLineNo">545</span><a name="line.545"></a> +<span class="sourceLineNo">546</span> /**<a name="line.546"></a> +<span class="sourceLineNo">547</span> * Returns sorted set of edit files made by splitter, excluding files<a name="line.547"></a> +<span class="sourceLineNo">548</span> * with '.temp' suffix.<a name="line.548"></a> +<span class="sourceLineNo">549</span> *<a name="line.549"></a> +<span class="sourceLineNo">550</span> * @param fs<a name="line.550"></a> +<span class="sourceLineNo">551</span> * @param regiondir<a name="line.551"></a> +<span class="sourceLineNo">552</span> * @return Files in passed <code>regiondir</code> as a sorted set.<a name="line.552"></a> +<span class="sourceLineNo">553</span> * @throws IOException<a name="line.553"></a> +<span class="sourceLineNo">554</span> */<a name="line.554"></a> +<span class="sourceLineNo">555</span> public static NavigableSet<Path> getSplitEditFilesSorted(final FileSystem fs,<a name="line.555"></a> +<span class="sourceLineNo">556</span> final Path regiondir) throws IOException {<a name="line.556"></a> +<span class="sourceLineNo">557</span> NavigableSet<Path> filesSorted = new TreeSet<>();<a name="line.557"></a> +<span class="sourceLineNo">558</span> Path editsdir = getRegionDirRecoveredEditsDir(regiondir);<a name="line.558"></a> +<span class="sourceLineNo">559</span> if (!fs.exists(editsdir)) {<a name="line.559"></a> +<span class="sourceLineNo">560</span> return filesSorted;<a name="line.560"></a> +<span class="sourceLineNo">561</span> }<a name="line.561"></a> +<span class="sourceLineNo">562</span> FileStatus[] files = FSUtils.listStatus(fs, editsdir, new PathFilter() {<a name="line.562"></a> +<span class="sourceLineNo">563</span> @Override<a name="line.563"></a> +<span class="sourceLineNo">564</span> public boolean accept(Path p) {<a name="line.564"></a> +<span class="sourceLineNo">565</span> boolean result = false;<a name="line.565"></a> +<span class="sourceLineNo">566</span> try {<a name="line.566"></a> +<span class="sourceLineNo">567</span> // Return files and only files that match the editfile names pattern.<a name="line.567"></a> +<span class="sourceLineNo">568</span> // There can be other files in this directory other than edit files.<a name="line.568"></a> +<span class="sourceLineNo">569</span> // In particular, on error, we'll move aside the bad edit file giving<a name="line.569"></a> +<span class="sourceLineNo">570</span> // it a timestamp suffix. See moveAsideBadEditsFile.<a name="line.570"></a> +<span class="sourceLineNo">571</span> Matcher m = EDITFILES_NAME_PATTERN.matcher(p.getName());<a name="line.571"></a> +<span class="sourceLineNo">572</span> result = fs.isFile(p) && m.matches();<a name="line.572"></a> +<span class="sourceLineNo">573</span> // Skip the file whose name ends with RECOVERED_LOG_TMPFILE_SUFFIX,<a name="line.573"></a> +<span class="sourceLineNo">574</span> // because it means splitwal thread is writting this file.<a name="line.574"></a> +<span class="sourceLineNo">575</span> if (p.getName().endsWith(RECOVERED_LOG_TMPFILE_SUFFIX)) {<a name="line.575"></a> +<span class="sourceLineNo">576</span> result = false;<a name="line.576"></a> +<span class="sourceLineNo">577</span> }<a name="line.577"></a> +<span class="sourceLineNo">578</span> // Skip SeqId Files<a name="line.578"></a> +<span class="sourceLineNo">579</span> if (isSequenceIdFile(p)) {<a name="line.579"></a> +<span class="sourceLineNo">580</span> result = false;<a name="line.580"></a> +<span class="sourceLineNo">581</span> }<a name="line.581"></a> +<span class="sourceLineNo">582</span> } catch (IOException e) {<a name="line.582"></a> +<span class="sourceLineNo">583</span> LOG.warn("Failed isFile check on {}", p, e);<a name="line.583"></a> +<span class="sourceLineNo">584</span> }<a name="line.584"></a> +<span class="sourceLineNo">585</span> return result;<a name="line.585"></a> +<span class="sourceLineNo">586</span> }<a name="line.586"></a> +<span class="sourceLineNo">587</span> });<a name="line.587"></a> +<span class="sourceLineNo">588</span> if (ArrayUtils.isNotEmpty(files)) {<a name="line.588"></a> +<span class="sourceLineNo">589</span> Arrays.asList(files).forEach(status -> filesSorted.add(status.getPath()));<a name="line.589"></a> +<span class="sourceLineNo">590</span> }<a name="line.590"></a> +<span class="sourceLineNo">591</span> return filesSorted;<a name="line.591"></a> +<span class="sourceLineNo">592</span> }<a name="line.592"></a> +<span class="sourceLineNo">593</span><a name="line.593"></a> +<span class="sourceLineNo">594</span> /**<a name="line.594"></a> +<span class="sourceLineNo">595</span> * Move aside a bad edits file.<a name="line.595"></a> +<span class="sourceLineNo">596</span> *<a name="line.596"></a> +<span class="sourceLineNo">597</span> * @param fs<a name="line.597"></a> +<span class="sourceLineNo">598</span> * @param edits<a name="line.598"></a> +<span class="sourceLineNo">599</span> * Edits file to move aside.<a name="line.599"></a> +<span class="sourceLineNo">600</span> * @return The name of the moved aside file.<a name="line.600"></a> +<span class="sourceLineNo">601</span> * @throws IOException<a name="line.601"></a> +<span class="sourceLineNo">602</span> */<a name="line.602"></a> +<span class="sourceLineNo">603</span> public static Path moveAsideBadEditsFile(final FileSystem fs, final Path edits)<a name="line.603"></a> +<span class="sourceLineNo">604</span> throws IOException {<a name="line.604"></a> +<span class="sourceLineNo">605</span> Path moveAsideName = new Path(edits.getParent(), edits.getName() + "."<a name="line.605"></a> +<span class="sourceLineNo">606</span> + System.currentTimeMillis());<a name="line.606"></a> +<span class="sourceLineNo">607</span> if (!fs.rename(edits, moveAsideName)) {<a name="line.607"></a> +<span class="sourceLineNo">608</span> LOG.warn("Rename failed from {} to {}", edits, moveAsideName);<a name="line.608"></a> +<span class="sourceLineNo">609</span> }<a name="line.609"></a> +<span class="sourceLineNo">610</span> return moveAsideName;<a name="line.610"></a> +<span class="sourceLineNo">611</span> }<a name="line.611"></a> +<span class="sourceLineNo">612</span><a name="line.612"></a> +<span class="sourceLineNo">613</span> private static final String SEQUENCE_ID_FILE_SUFFIX = ".seqid";<a name="line.613"></a> +<span class="sourceLineNo">614</span> private static final String OLD_SEQUENCE_ID_FILE_SUFFIX = "_seqid";<a name="line.614"></a> +<span class="sourceLineNo">615</span> private static final int SEQUENCE_ID_FILE_SUFFIX_LENGTH = SEQUENCE_ID_FILE_SUFFIX.length();<a name="line.615"></a> +<span class="sourceLineNo">616</span><a name="line.616"></a> +<span class="sourceLineNo">617</span> /**<a name="line.617"></a> +<span class="sourceLineNo">618</span> * Is the given file a region open sequence id file.<a name="line.618"></a> +<span class="sourceLineNo">619</span> */<a name="line.619"></a> +<span class="sourceLineNo">620</span> @VisibleForTesting<a name="line.620"></a> +<span class="sourceLineNo">621</span> public static boolean isSequenceIdFile(final Path file) {<a name="line.621"></a> +<span class="sourceLineNo">622</span> return file.getName().endsWith(SEQUENCE_ID_FILE_SUFFIX)<a name="line.622"></a> +<span class="sourceLineNo">623</span> || file.getName().endsWith(OLD_SEQUENCE_ID_FILE_SUFFIX);<a name="line.623"></a> +<span class="sourceLineNo">624</span> }<a name="line.624"></a> +<span class="sourceLineNo">625</span><a name="line.625"></a> +<span class="sourceLineNo">626</span> private static FileStatus[] getSequenceIdFiles(FileSystem fs, Path regionDir) throws IOException {<a name="line.626"></a> +<span class="sourceLineNo">627</span> // TODO: Why are we using a method in here as part of our normal region open where<a name="line.627"></a> +<span class="sourceLineNo">628</span> // there is no splitting involved? Fix. St.Ack 01/20/2017.<a name="line.628"></a> +<span class="sourceLineNo">629</span> Path editsDir = WALSplitter.getRegionDirRecoveredEditsDir(regionDir);<a name="line.629"></a> +<span class="sourceLineNo">630</span> try {<a name="line.630"></a> +<span class="sourceLineNo">631</span> FileStatus[] files = fs.listStatus(editsDir, WALSplitter::isSequenceIdFile);<a name="line.631"></a> +<span class="sourceLineNo">632</span> return files != null ? files : new FileStatus[0];<a name="line.632"></a> +<span class="sourceLineNo">633</span> } catch (FileNotFoundException e) {<a name="line.633"></a> +<span class="sourceLineNo">634</span> return new FileStatus[0];<a name="line.634"></a> +<span class="sourceLineNo">635</span> }<a name="line.635"></a> +<span class="sourceLineNo">636</span> }<a name="line.636"></a> +<span class="sourceLineNo">637</span><a name="line.637"></a> +<span class="sourceLineNo">638</span> private static long getMaxSequenceId(FileStatus[] files) {<a name="line.638"></a> +<span class="sourceLineNo">639</span> long maxSeqId = -1L;<a name="line.639"></a> +<span class="sourceLineNo">640</span> for (FileStatus file : files) {<a name="line.640"></a> +<span class="sourceLineNo">641</span> String fileName = file.getPath().getName();<a name="line.641"></a> +<span class="sourceLineNo">642</span> try {<a name="line.642"></a> +<span class="sourceLineNo">643</span> maxSeqId = Math.max(maxSeqId, Long<a name="line.643"></a> +<span class="sourceLineNo">644</span> .parseLong(fileName.substring(0, fileName.length() - SEQUENCE_ID_FILE_SUFFIX_LENGTH)));<a name="line.644"></a> +<span class="sourceLineNo">645</span> } catch (NumberFormatException ex) {<a name="line.645"></a> +<span class="sourceLineNo">646</span> LOG.warn("Invalid SeqId File Name={}", fileName);<a name="line.646"></a> +<span class="sourceLineNo">647</span> }<a name="line.647"></a> +<span class="sourceLineNo">648</span> }<a name="line.648"></a> +<span class="sourceLineNo">649</span> return maxSeqId;<a name="line.649"></a> +<span class="sourceLineNo">650</span> }<a name="line.650"></a> +<span class="sourceLineNo">651</span><a name="line.651"></a> +<span class="sourceLineNo">652</span> /**<a name="line.652"></a> +<span class="sourceLineNo">653</span> * Get the max sequence id which is stored in the region directory. -1 if none.<a name="line.653"></a> +<span class="sourceLineNo">654</span> */<a name="line.654"></a> +<span class="sourceLineNo">655</span> public static long getMaxRegionSequenceId(FileSystem fs, Path regionDir) throws IOException {<a name="line.655"></a> +<span class="sourceLineNo">656</span> return getMaxSequenceId(getSequenceIdFiles(fs, regionDir));<a name="line.656"></a> +<span class="sourceLineNo">657</span> }<a name="line.657"></a> +<span class="sourceLineNo">658</span><a name="line.658"></a> +<span class="sourceLineNo">659</span> /**<a name="line.659"></a> +<span class="sourceLineNo">660</span> * Create a file with name as region's max sequence id<a name="line.660"></a> +<span class="sourceLineNo">661</span> */<a name="line.661"></a> +<span class="sourceLineNo">662</span> public static void writeRegionSequenceIdFile(FileSystem fs, Path regionDir, long newMaxSeqId)<a name="line.662"></a> +<span class="sourceLineNo">663</span> throws IOException {<a name="line.663"></a> +<span class="sourceLineNo">664</span> FileStatus[] files = getSequenceIdFiles(fs, regionDir);<a name="line.664"></a> +<span class="sourceLineNo">665</span> long maxSeqId = getMaxSequenceId(files);<a name="line.665"></a> +<span class="sourceLineNo">666</span> if (maxSeqId > newMaxSeqId) {<a name="line.666"></a> +<span class="sourceLineNo">667</span> throw new IOException("The new max sequence id " + newMaxSeqId +<a name="line.667"></a> +<span class="sourceLineNo">668</span> " is less than the old max sequence id " + maxSeqId);<a name="line.668"></a> +<span class="sourceLineNo">669</span> }<a name="line.669"></a> +<span class="sourceLineNo">670</span> // write a new seqId file<a name="line.670"></a> +<span class="sourceLineNo">671</span> Path newSeqIdFile = new Path(WALSplitter.getRegionDirRecoveredEditsDir(regionDir),<a name="line.671"></a> +<span class="sourceLineNo">672</span> newMaxSeqId + SEQUENCE_ID_FILE_SUFFIX);<a name="line.672"></a> +<span class="sourceLineNo">673</span> if (newMaxSeqId != maxSeqId) {<a name="line.673"></a> +<span class="sourceLineNo">674</span> try {<a name="line.674"></a> +<span class="sourceLineNo">675</span> if (!fs.createNewFile(newSeqIdFile) && !fs.exists(newSeqIdFile)) {<a name="line.675"></a> +<span class="sourceLineNo">676</span> throw new IOException("Failed to create SeqId file:" + newSeqIdFile);<a name="line.676"></a> +<span class="sourceLineNo">677</span> }<a name="line.677"></a> +<span class="sourceLineNo">678</span> LOG.debug("Wrote file={}, newMaxSeqId={}, maxSeqId={}", newSeqIdFile, newMaxSeqId,<a name="line.678"></a> +<span class="sourceLineNo">679</span> maxSeqId);<a name="line.679"></a> +<span class="sourceLineNo">680</span> } catch (FileAlreadyExistsException ignored) {<a name="line.680"></a> +<span class="sourceLineNo">681</span> // latest hdfs throws this exception. it's all right if newSeqIdFile already exists<a name="line.681"></a> +<span class="sourceL
<TRUNCATED>
