You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hbase.apache.org by mi...@apache.org on 2015/12/18 17:43:00 UTC
[27/48] hbase-site git commit: Published site at
4bfeccb87a94cfe232ea8fc9a6f40ff5b8d3b1c5.
http://git-wip-us.apache.org/repos/asf/hbase-site/blob/d917c66a/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html
----------------------------------------------------------------------
diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html b/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html
index 5952d6a..7c00453 100644
--- a/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html
+++ b/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html
@@ -433,654 +433,655 @@
<span class="sourceLineNo">425</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.425"></a>
<span class="sourceLineNo">426</span> * @param table table to which hfiles should be loaded<a name="line.426"></a>
<span class="sourceLineNo">427</span> * @param queue queue which needs to be loaded into the table<a name="line.427"></a>
-<span class="sourceLineNo">428</span> * @throws IOException If any I/O or network error occurred<a name="line.428"></a>
-<span class="sourceLineNo">429</span> */<a name="line.429"></a>
-<span class="sourceLineNo">430</span> public void prepareHFileQueue(Path hfofDir, Table table, Deque<LoadQueueItem> queue,<a name="line.430"></a>
-<span class="sourceLineNo">431</span> boolean validateHFile) throws IOException {<a name="line.431"></a>
-<span class="sourceLineNo">432</span> discoverLoadQueue(queue, hfofDir, validateHFile);<a name="line.432"></a>
-<span class="sourceLineNo">433</span> validateFamiliesInHFiles(table, queue);<a name="line.433"></a>
-<span class="sourceLineNo">434</span> }<a name="line.434"></a>
-<span class="sourceLineNo">435</span><a name="line.435"></a>
-<span class="sourceLineNo">436</span> // Initialize a thread pool<a name="line.436"></a>
-<span class="sourceLineNo">437</span> private ExecutorService createExecutorService() {<a name="line.437"></a>
-<span class="sourceLineNo">438</span> ThreadFactoryBuilder builder = new ThreadFactoryBuilder();<a name="line.438"></a>
-<span class="sourceLineNo">439</span> builder.setNameFormat("LoadIncrementalHFiles-%1$d");<a name="line.439"></a>
-<span class="sourceLineNo">440</span> ExecutorService pool = new ThreadPoolExecutor(nrThreads, nrThreads, 60, TimeUnit.SECONDS,<a name="line.440"></a>
-<span class="sourceLineNo">441</span> new LinkedBlockingQueue<Runnable>(), builder.build());<a name="line.441"></a>
-<span class="sourceLineNo">442</span> ((ThreadPoolExecutor) pool).allowCoreThreadTimeOut(true);<a name="line.442"></a>
-<span class="sourceLineNo">443</span> return pool;<a name="line.443"></a>
-<span class="sourceLineNo">444</span> }<a name="line.444"></a>
-<span class="sourceLineNo">445</span><a name="line.445"></a>
-<span class="sourceLineNo">446</span> /**<a name="line.446"></a>
-<span class="sourceLineNo">447</span> * Checks whether there is any invalid family name in HFiles to be bulk loaded.<a name="line.447"></a>
-<span class="sourceLineNo">448</span> */<a name="line.448"></a>
-<span class="sourceLineNo">449</span> private void validateFamiliesInHFiles(Table table, Deque<LoadQueueItem> queue)<a name="line.449"></a>
-<span class="sourceLineNo">450</span> throws IOException {<a name="line.450"></a>
-<span class="sourceLineNo">451</span> Collection<HColumnDescriptor> families = table.getTableDescriptor().getFamilies();<a name="line.451"></a>
-<span class="sourceLineNo">452</span> List<String> familyNames = new ArrayList<String>(families.size());<a name="line.452"></a>
-<span class="sourceLineNo">453</span> for (HColumnDescriptor family : families) {<a name="line.453"></a>
-<span class="sourceLineNo">454</span> familyNames.add(family.getNameAsString());<a name="line.454"></a>
-<span class="sourceLineNo">455</span> }<a name="line.455"></a>
-<span class="sourceLineNo">456</span> List<String> unmatchedFamilies = new ArrayList<String>();<a name="line.456"></a>
-<span class="sourceLineNo">457</span> Iterator<LoadQueueItem> queueIter = queue.iterator();<a name="line.457"></a>
-<span class="sourceLineNo">458</span> while (queueIter.hasNext()) {<a name="line.458"></a>
-<span class="sourceLineNo">459</span> LoadQueueItem lqi = queueIter.next();<a name="line.459"></a>
-<span class="sourceLineNo">460</span> String familyNameInHFile = Bytes.toString(lqi.family);<a name="line.460"></a>
-<span class="sourceLineNo">461</span> if (!familyNames.contains(familyNameInHFile)) {<a name="line.461"></a>
-<span class="sourceLineNo">462</span> unmatchedFamilies.add(familyNameInHFile);<a name="line.462"></a>
-<span class="sourceLineNo">463</span> }<a name="line.463"></a>
-<span class="sourceLineNo">464</span> }<a name="line.464"></a>
-<span class="sourceLineNo">465</span> if (unmatchedFamilies.size() > 0) {<a name="line.465"></a>
-<span class="sourceLineNo">466</span> String msg =<a name="line.466"></a>
-<span class="sourceLineNo">467</span> "Unmatched family names found: unmatched family names in HFiles to be bulkloaded: "<a name="line.467"></a>
-<span class="sourceLineNo">468</span> + unmatchedFamilies + "; valid family names of table " + table.getName() + " are: "<a name="line.468"></a>
-<span class="sourceLineNo">469</span> + familyNames;<a name="line.469"></a>
-<span class="sourceLineNo">470</span> LOG.error(msg);<a name="line.470"></a>
-<span class="sourceLineNo">471</span> throw new IOException(msg);<a name="line.471"></a>
-<span class="sourceLineNo">472</span> }<a name="line.472"></a>
-<span class="sourceLineNo">473</span> }<a name="line.473"></a>
-<span class="sourceLineNo">474</span><a name="line.474"></a>
-<span class="sourceLineNo">475</span> /**<a name="line.475"></a>
-<span class="sourceLineNo">476</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.476"></a>
-<span class="sourceLineNo">477</span> * 1. {@link LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)} 2.<a name="line.477"></a>
-<span class="sourceLineNo">478</span> * {@link<a name="line.478"></a>
-<span class="sourceLineNo">479</span> * LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)}<a name="line.479"></a>
-<span class="sourceLineNo">480</span> * @param table Table to which these hfiles should be loaded to<a name="line.480"></a>
-<span class="sourceLineNo">481</span> * @param conn Connection to use<a name="line.481"></a>
-<span class="sourceLineNo">482</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.482"></a>
-<span class="sourceLineNo">483</span> * @param startEndKeys starting and ending row keys of the region<a name="line.483"></a>
-<span class="sourceLineNo">484</span> */<a name="line.484"></a>
-<span class="sourceLineNo">485</span> public void loadHFileQueue(final Table table, final Connection conn, Deque<LoadQueueItem> queue,<a name="line.485"></a>
-<span class="sourceLineNo">486</span> Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.486"></a>
-<span class="sourceLineNo">487</span> ExecutorService pool = null;<a name="line.487"></a>
-<span class="sourceLineNo">488</span> try {<a name="line.488"></a>
-<span class="sourceLineNo">489</span> pool = createExecutorService();<a name="line.489"></a>
-<span class="sourceLineNo">490</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups =<a name="line.490"></a>
-<span class="sourceLineNo">491</span> groupOrSplitPhase(table, pool, queue, startEndKeys);<a name="line.491"></a>
-<span class="sourceLineNo">492</span> bulkLoadPhase(table, conn, pool, queue, regionGroups);<a name="line.492"></a>
-<span class="sourceLineNo">493</span> } finally {<a name="line.493"></a>
-<span class="sourceLineNo">494</span> if (pool != null) {<a name="line.494"></a>
-<span class="sourceLineNo">495</span> pool.shutdown();<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><a name="line.499"></a>
-<span class="sourceLineNo">500</span> /**<a name="line.500"></a>
-<span class="sourceLineNo">501</span> * This takes the LQI's grouped by likely regions and attempts to bulk load<a name="line.501"></a>
-<span class="sourceLineNo">502</span> * them. Any failures are re-queued for another pass with the<a name="line.502"></a>
-<span class="sourceLineNo">503</span> * groupOrSplitPhase.<a name="line.503"></a>
-<span class="sourceLineNo">504</span> */<a name="line.504"></a>
-<span class="sourceLineNo">505</span> protected void bulkLoadPhase(final Table table, final Connection conn,<a name="line.505"></a>
-<span class="sourceLineNo">506</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.506"></a>
-<span class="sourceLineNo">507</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) throws IOException {<a name="line.507"></a>
-<span class="sourceLineNo">508</span> // atomically bulk load the groups.<a name="line.508"></a>
-<span class="sourceLineNo">509</span> Set<Future<List<LoadQueueItem>>> loadingFutures = new HashSet<Future<List<LoadQueueItem>>>();<a name="line.509"></a>
-<span class="sourceLineNo">510</span> for (Entry<ByteBuffer, ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()){<a name="line.510"></a>
-<span class="sourceLineNo">511</span> final byte[] first = e.getKey().array();<a name="line.511"></a>
-<span class="sourceLineNo">512</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.512"></a>
-<span class="sourceLineNo">513</span><a name="line.513"></a>
-<span class="sourceLineNo">514</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.514"></a>
-<span class="sourceLineNo">515</span> @Override<a name="line.515"></a>
-<span class="sourceLineNo">516</span> public List<LoadQueueItem> call() throws Exception {<a name="line.516"></a>
-<span class="sourceLineNo">517</span> List<LoadQueueItem> toRetry =<a name="line.517"></a>
-<span class="sourceLineNo">518</span> tryAtomicRegionLoad(conn, table.getName(), first, lqis);<a name="line.518"></a>
-<span class="sourceLineNo">519</span> return toRetry;<a name="line.519"></a>
-<span class="sourceLineNo">520</span> }<a name="line.520"></a>
-<span class="sourceLineNo">521</span> };<a name="line.521"></a>
-<span class="sourceLineNo">522</span> loadingFutures.add(pool.submit(call));<a name="line.522"></a>
-<span class="sourceLineNo">523</span> }<a name="line.523"></a>
-<span class="sourceLineNo">524</span><a name="line.524"></a>
-<span class="sourceLineNo">525</span> // get all the results.<a name="line.525"></a>
-<span class="sourceLineNo">526</span> for (Future<List<LoadQueueItem>> future : loadingFutures) {<a name="line.526"></a>
-<span class="sourceLineNo">527</span> try {<a name="line.527"></a>
-<span class="sourceLineNo">528</span> List<LoadQueueItem> toRetry = future.get();<a name="line.528"></a>
-<span class="sourceLineNo">529</span><a name="line.529"></a>
-<span class="sourceLineNo">530</span> // LQIs that are requeued to be regrouped.<a name="line.530"></a>
-<span class="sourceLineNo">531</span> queue.addAll(toRetry);<a name="line.531"></a>
-<span class="sourceLineNo">532</span><a name="line.532"></a>
-<span class="sourceLineNo">533</span> } catch (ExecutionException e1) {<a name="line.533"></a>
-<span class="sourceLineNo">534</span> Throwable t = e1.getCause();<a name="line.534"></a>
-<span class="sourceLineNo">535</span> if (t instanceof IOException) {<a name="line.535"></a>
-<span class="sourceLineNo">536</span> // At this point something unrecoverable has happened.<a name="line.536"></a>
-<span class="sourceLineNo">537</span> // TODO Implement bulk load recovery<a name="line.537"></a>
-<span class="sourceLineNo">538</span> throw new IOException("BulkLoad encountered an unrecoverable problem", t);<a name="line.538"></a>
-<span class="sourceLineNo">539</span> }<a name="line.539"></a>
-<span class="sourceLineNo">540</span> LOG.error("Unexpected execution exception during bulk load", e1);<a name="line.540"></a>
-<span class="sourceLineNo">541</span> throw new IllegalStateException(t);<a name="line.541"></a>
-<span class="sourceLineNo">542</span> } catch (InterruptedException e1) {<a name="line.542"></a>
-<span class="sourceLineNo">543</span> LOG.error("Unexpected interrupted exception during bulk load", e1);<a name="line.543"></a>
-<span class="sourceLineNo">544</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<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> }<a name="line.547"></a>
-<span class="sourceLineNo">548</span><a name="line.548"></a>
-<span class="sourceLineNo">549</span> private boolean checkHFilesCountPerRegionPerFamily(<a name="line.549"></a>
-<span class="sourceLineNo">550</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) {<a name="line.550"></a>
-<span class="sourceLineNo">551</span> for (Entry<ByteBuffer,<a name="line.551"></a>
-<span class="sourceLineNo">552</span> ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()) {<a name="line.552"></a>
-<span class="sourceLineNo">553</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.553"></a>
-<span class="sourceLineNo">554</span> HashMap<byte[], MutableInt> filesMap = new HashMap<byte[], MutableInt>();<a name="line.554"></a>
-<span class="sourceLineNo">555</span> for (LoadQueueItem lqi: lqis) {<a name="line.555"></a>
-<span class="sourceLineNo">556</span> MutableInt count = filesMap.get(lqi.family);<a name="line.556"></a>
-<span class="sourceLineNo">557</span> if (count == null) {<a name="line.557"></a>
-<span class="sourceLineNo">558</span> count = new MutableInt();<a name="line.558"></a>
-<span class="sourceLineNo">559</span> filesMap.put(lqi.family, count);<a name="line.559"></a>
-<span class="sourceLineNo">560</span> }<a name="line.560"></a>
-<span class="sourceLineNo">561</span> count.increment();<a name="line.561"></a>
-<span class="sourceLineNo">562</span> if (count.intValue() > maxFilesPerRegionPerFamily) {<a name="line.562"></a>
-<span class="sourceLineNo">563</span> LOG.error("Trying to load more than " + maxFilesPerRegionPerFamily<a name="line.563"></a>
-<span class="sourceLineNo">564</span> + " hfiles to family " + Bytes.toStringBinary(lqi.family)<a name="line.564"></a>
-<span class="sourceLineNo">565</span> + " of region with start key "<a name="line.565"></a>
-<span class="sourceLineNo">566</span> + Bytes.toStringBinary(e.getKey()));<a name="line.566"></a>
-<span class="sourceLineNo">567</span> return false;<a name="line.567"></a>
-<span class="sourceLineNo">568</span> }<a name="line.568"></a>
-<span class="sourceLineNo">569</span> }<a name="line.569"></a>
-<span class="sourceLineNo">570</span> }<a name="line.570"></a>
-<span class="sourceLineNo">571</span> return true;<a name="line.571"></a>
-<span class="sourceLineNo">572</span> }<a name="line.572"></a>
-<span class="sourceLineNo">573</span><a name="line.573"></a>
-<span class="sourceLineNo">574</span> /**<a name="line.574"></a>
-<span class="sourceLineNo">575</span> * @return A map that groups LQI by likely bulk load region targets.<a name="line.575"></a>
-<span class="sourceLineNo">576</span> */<a name="line.576"></a>
-<span class="sourceLineNo">577</span> private Multimap<ByteBuffer, LoadQueueItem> groupOrSplitPhase(final Table table,<a name="line.577"></a>
-<span class="sourceLineNo">578</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.578"></a>
-<span class="sourceLineNo">579</span> final Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.579"></a>
-<span class="sourceLineNo">580</span> // <region start key, LQI> need synchronized only within this scope of this<a name="line.580"></a>
-<span class="sourceLineNo">581</span> // phase because of the puts that happen in futures.<a name="line.581"></a>
-<span class="sourceLineNo">582</span> Multimap<ByteBuffer, LoadQueueItem> rgs = HashMultimap.create();<a name="line.582"></a>
-<span class="sourceLineNo">583</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups = Multimaps.synchronizedMultimap(rgs);<a name="line.583"></a>
-<span class="sourceLineNo">584</span><a name="line.584"></a>
-<span class="sourceLineNo">585</span> // drain LQIs and figure out bulk load groups<a name="line.585"></a>
-<span class="sourceLineNo">586</span> Set<Future<List<LoadQueueItem>>> splittingFutures = new HashSet<Future<List<LoadQueueItem>>>();<a name="line.586"></a>
-<span class="sourceLineNo">587</span> while (!queue.isEmpty()) {<a name="line.587"></a>
-<span class="sourceLineNo">588</span> final LoadQueueItem item = queue.remove();<a name="line.588"></a>
-<span class="sourceLineNo">589</span><a name="line.589"></a>
-<span class="sourceLineNo">590</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.590"></a>
-<span class="sourceLineNo">591</span> @Override<a name="line.591"></a>
-<span class="sourceLineNo">592</span> public List<LoadQueueItem> call() throws Exception {<a name="line.592"></a>
-<span class="sourceLineNo">593</span> List<LoadQueueItem> splits = groupOrSplit(regionGroups, item, table, startEndKeys);<a name="line.593"></a>
-<span class="sourceLineNo">594</span> return splits;<a name="line.594"></a>
-<span class="sourceLineNo">595</span> }<a name="line.595"></a>
-<span class="sourceLineNo">596</span> };<a name="line.596"></a>
-<span class="sourceLineNo">597</span> splittingFutures.add(pool.submit(call));<a name="line.597"></a>
-<span class="sourceLineNo">598</span> }<a name="line.598"></a>
-<span class="sourceLineNo">599</span> // get all the results. All grouping and splitting must finish before<a name="line.599"></a>
-<span class="sourceLineNo">600</span> // we can attempt the atomic loads.<a name="line.600"></a>
-<span class="sourceLineNo">601</span> for (Future<List<LoadQueueItem>> lqis : splittingFutures) {<a name="line.601"></a>
-<span class="sourceLineNo">602</span> try {<a name="line.602"></a>
-<span class="sourceLineNo">603</span> List<LoadQueueItem> splits = lqis.get();<a name="line.603"></a>
-<span class="sourceLineNo">604</span> if (splits != null) {<a name="line.604"></a>
-<span class="sourceLineNo">605</span> queue.addAll(splits);<a name="line.605"></a>
-<span class="sourceLineNo">606</span> }<a name="line.606"></a>
-<span class="sourceLineNo">607</span> } catch (ExecutionException e1) {<a name="line.607"></a>
-<span class="sourceLineNo">608</span> Throwable t = e1.getCause();<a name="line.608"></a>
-<span class="sourceLineNo">609</span> if (t instanceof IOException) {<a name="line.609"></a>
-<span class="sourceLineNo">610</span> LOG.error("IOException during splitting", e1);<a name="line.610"></a>
-<span class="sourceLineNo">611</span> throw (IOException)t; // would have been thrown if not parallelized,<a name="line.611"></a>
-<span class="sourceLineNo">612</span> }<a name="line.612"></a>
-<span class="sourceLineNo">613</span> LOG.error("Unexpected execution exception during splitting", e1);<a name="line.613"></a>
-<span class="sourceLineNo">614</span> throw new IllegalStateException(t);<a name="line.614"></a>
-<span class="sourceLineNo">615</span> } catch (InterruptedException e1) {<a name="line.615"></a>
-<span class="sourceLineNo">616</span> LOG.error("Unexpected interrupted exception during splitting", e1);<a name="line.616"></a>
-<span class="sourceLineNo">617</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<a name="line.617"></a>
-<span class="sourceLineNo">618</span> }<a name="line.618"></a>
-<span class="sourceLineNo">619</span> }<a name="line.619"></a>
-<span class="sourceLineNo">620</span> return regionGroups;<a name="line.620"></a>
-<span class="sourceLineNo">621</span> }<a name="line.621"></a>
-<span class="sourceLineNo">622</span><a name="line.622"></a>
-<span class="sourceLineNo">623</span> // unique file name for the table<a name="line.623"></a>
-<span class="sourceLineNo">624</span> private String getUniqueName() {<a name="line.624"></a>
-<span class="sourceLineNo">625</span> return UUID.randomUUID().toString().replaceAll("-", "");<a name="line.625"></a>
-<span class="sourceLineNo">626</span> }<a name="line.626"></a>
-<span class="sourceLineNo">627</span><a name="line.627"></a>
-<span class="sourceLineNo">628</span> protected List<LoadQueueItem> splitStoreFile(final LoadQueueItem item,<a name="line.628"></a>
-<span class="sourceLineNo">629</span> final Table table, byte[] startKey,<a name="line.629"></a>
-<span class="sourceLineNo">630</span> byte[] splitKey) throws IOException {<a name="line.630"></a>
-<span class="sourceLineNo">631</span> final Path hfilePath = item.hfilePath;<a name="line.631"></a>
-<span class="sourceLineNo">632</span><a name="line.632"></a>
-<span class="sourceLineNo">633</span> // We use a '_' prefix which is ignored when walking directory trees<a name="line.633"></a>
-<span class="sourceLineNo">634</span> // above.<a name="line.634"></a>
-<span class="sourceLineNo">635</span> final String TMP_DIR = "_tmp";<a name="line.635"></a>
-<span class="sourceLineNo">636</span> Path tmpDir = item.hfilePath.getParent();<a name="line.636"></a>
-<span class="sourceLineNo">637</span> if (!tmpDir.getName().equals(TMP_DIR)) {<a name="line.637"></a>
-<span class="sourceLineNo">638</span> tmpDir = new Path(tmpDir, TMP_DIR);<a name="line.638"></a>
-<span class="sourceLineNo">639</span> }<a name="line.639"></a>
-<span class="sourceLineNo">640</span><a name="line.640"></a>
-<span class="sourceLineNo">641</span> LOG.info("HFile at " + hfilePath + " no longer fits inside a single " +<a name="line.641"></a>
-<span class="sourceLineNo">642</span> "region. Splitting...");<a name="line.642"></a>
-<span class="sourceLineNo">643</span><a name="line.643"></a>
-<span class="sourceLineNo">644</span> String uniqueName = getUniqueName();<a name="line.644"></a>
-<span class="sourceLineNo">645</span> HColumnDescriptor familyDesc = table.getTableDescriptor().getFamily(item.family);<a name="line.645"></a>
-<span class="sourceLineNo">646</span><a name="line.646"></a>
-<span class="sourceLineNo">647</span> Path botOut = new Path(tmpDir, uniqueName + ".bottom");<a name="line.647"></a>
-<span class="sourceLineNo">648</span> Path topOut = new Path(tmpDir, uniqueName + ".top");<a name="line.648"></a>
-<span class="sourceLineNo">649</span> splitStoreFile(getConf(), hfilePath, familyDesc, splitKey, botOut, topOut);<a name="line.649"></a>
-<span class="sourceLineNo">650</span><a name="line.650"></a>
-<span class="sourceLineNo">651</span> FileSystem fs = tmpDir.getFileSystem(getConf());<a name="line.651"></a>
-<span class="sourceLineNo">652</span> fs.setPermission(tmpDir, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.652"></a>
-<span class="sourceLineNo">653</span> fs.setPermission(botOut, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.653"></a>
-<span class="sourceLineNo">654</span> fs.setPermission(topOut, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.654"></a>
-<span class="sourceLineNo">655</span><a name="line.655"></a>
-<span class="sourceLineNo">656</span> // Add these back at the *front* of the queue, so there's a lower<a name="line.656"></a>
-<span class="sourceLineNo">657</span> // chance that the region will just split again before we get there.<a name="line.657"></a>
-<span class="sourceLineNo">658</span> List<LoadQueueItem> lqis = new ArrayList<LoadQueueItem>(2);<a name="line.658"></a>
-<span class="sourceLineNo">659</span> lqis.add(new LoadQueueItem(item.family, botOut));<a name="line.659"></a>
-<span class="sourceLineNo">660</span> lqis.add(new LoadQueueItem(item.family, topOut));<a name="line.660"></a>
-<span class="sourceLineNo">661</span><a name="line.661"></a>
-<span class="sourceLineNo">662</span> LOG.info("Successfully split into new HFiles " + botOut + " and " + topOut);<a name="line.662"></a>
-<span class="sourceLineNo">663</span> return lqis;<a name="line.663"></a>
-<span class="sourceLineNo">664</span> }<a name="line.664"></a>
-<span class="sourceLineNo">665</span><a name="line.665"></a>
-<span class="sourceLineNo">666</span> /**<a name="line.666"></a>
-<span class="sourceLineNo">667</span> * Attempt to assign the given load queue item into its target region group.<a name="line.667"></a>
-<span class="sourceLineNo">668</span> * If the hfile boundary no longer fits into a region, physically splits<a name="line.668"></a>
-<span class="sourceLineNo">669</span> * the hfile such that the new bottom half will fit and returns the list of<a name="line.669"></a>
-<span class="sourceLineNo">670</span> * LQI's corresponding to the resultant hfiles.<a name="line.670"></a>
-<span class="sourceLineNo">671</span> *<a name="line.671"></a>
-<span class="sourceLineNo">672</span> * protected for testing<a name="line.672"></a>
-<span class="sourceLineNo">673</span> * @throws IOException<a name="line.673"></a>
-<span class="sourceLineNo">674</span> */<a name="line.674"></a>
-<span class="sourceLineNo">675</span> protected List<LoadQueueItem> groupOrSplit(Multimap<ByteBuffer, LoadQueueItem> regionGroups,<a name="line.675"></a>
-<span class="sourceLineNo">676</span> final LoadQueueItem item, final Table table,<a name="line.676"></a>
-<span class="sourceLineNo">677</span> final Pair<byte[][], byte[][]> startEndKeys)<a name="line.677"></a>
-<span class="sourceLineNo">678</span> throws IOException {<a name="line.678"></a>
-<span class="sourceLineNo">679</span> final Path hfilePath = item.hfilePath;<a name="line.679"></a>
-<span class="sourceLineNo">680</span> // fs is the source filesystem<a name="line.680"></a>
-<span class="sourceLineNo">681</span> if (fs == null) {<a name="line.681"></a>
-<span class="sourceLineNo">682</span> fs = hfilePath.getFileSystem(getConf());<a name="line.682"></a>
-<span class="sourceLineNo">683</span> }<a name="line.683"></a>
-<span class="sourceLineNo">684</span> HFile.Reader hfr = HFile.createReader(fs, hfilePath,<a name="line.684"></a>
-<span class="sourceLineNo">685</span> new CacheConfig(getConf()), getConf());<a name="line.685"></a>
-<span class="sourceLineNo">686</span> final byte[] first, last;<a name="line.686"></a>
-<span class="sourceLineNo">687</span> try {<a name="line.687"></a>
-<span class="sourceLineNo">688</span> hfr.loadFileInfo();<a name="line.688"></a>
-<span class="sourceLineNo">689</span> first = hfr.getFirstRowKey();<a name="line.689"></a>
-<span class="sourceLineNo">690</span> last = hfr.getLastRowKey();<a name="line.690"></a>
-<span class="sourceLineNo">691</span> } finally {<a name="line.691"></a>
-<span class="sourceLineNo">692</span> hfr.close();<a name="line.692"></a>
-<span class="sourceLineNo">693</span> }<a name="line.693"></a>
-<span class="sourceLineNo">694</span><a name="line.694"></a>
-<span class="sourceLineNo">695</span> LOG.info("Trying to load hfile=" + hfilePath +<a name="line.695"></a>
-<span class="sourceLineNo">696</span> " first=" + Bytes.toStringBinary(first) +<a name="line.696"></a>
-<span class="sourceLineNo">697</span> " last=" + Bytes.toStringBinary(last));<a name="line.697"></a>
-<span class="sourceLineNo">698</span> if (first == null || last == null) {<a name="line.698"></a>
-<span class="sourceLineNo">699</span> assert first == null && last == null;<a name="line.699"></a>
-<span class="sourceLineNo">700</span> // TODO what if this is due to a bad HFile?<a name="line.700"></a>
-<span class="sourceLineNo">701</span> LOG.info("hfile " + hfilePath + " has no entries, skipping");<a name="line.701"></a>
-<span class="sourceLineNo">702</span> return null;<a name="line.702"></a>
-<span class="sourceLineNo">703</span> }<a name="line.703"></a>
-<span class="sourceLineNo">704</span> if (Bytes.compareTo(first, last) > 0) {<a name="line.704"></a>
-<span class="sourceLineNo">705</span> throw new IllegalArgumentException(<a name="line.705"></a>
-<span class="sourceLineNo">706</span> "Invalid range: " + Bytes.toStringBinary(first) +<a name="line.706"></a>
-<span class="sourceLineNo">707</span> " > " + Bytes.toStringBinary(last));<a name="line.707"></a>
-<span class="sourceLineNo">708</span> }<a name="line.708"></a>
-<span class="sourceLineNo">709</span> int idx = Arrays.binarySearch(startEndKeys.getFirst(), first,<a name="line.709"></a>
-<span class="sourceLineNo">710</span> Bytes.BYTES_COMPARATOR);<a name="line.710"></a>
-<span class="sourceLineNo">711</span> if (idx < 0) {<a name="line.711"></a>
-<span class="sourceLineNo">712</span> // not on boundary, returns -(insertion index). Calculate region it<a name="line.712"></a>
-<span class="sourceLineNo">713</span> // would be in.<a name="line.713"></a>
-<span class="sourceLineNo">714</span> idx = -(idx + 1) - 1;<a name="line.714"></a>
-<span class="sourceLineNo">715</span> }<a name="line.715"></a>
-<span class="sourceLineNo">716</span> final int indexForCallable = idx;<a name="line.716"></a>
-<span class="sourceLineNo">717</span><a name="line.717"></a>
-<span class="sourceLineNo">718</span> /**<a name="line.718"></a>
-<span class="sourceLineNo">719</span> * we can consider there is a region hole in following conditions. 1) if idx < 0,then first<a name="line.719"></a>
-<span class="sourceLineNo">720</span> * region info is lost. 2) if the endkey of a region is not equal to the startkey of the next<a name="line.720"></a>
-<span class="sourceLineNo">721</span> * region. 3) if the endkey of the last region is not empty.<a name="line.721"></a>
-<span class="sourceLineNo">722</span> */<a name="line.722"></a>
-<span class="sourceLineNo">723</span> if (indexForCallable < 0) {<a name="line.723"></a>
-<span class="sourceLineNo">724</span> throw new IOException("The first region info for table "<a name="line.724"></a>
-<span class="sourceLineNo">725</span> + table.getName()<a name="line.725"></a>
-<span class="sourceLineNo">726</span> + " cann't be found in hbase:meta.Please use hbck tool to fix it first.");<a name="line.726"></a>
-<span class="sourceLineNo">727</span> } else if ((indexForCallable == startEndKeys.getFirst().length - 1)<a name="line.727"></a>
-<span class="sourceLineNo">728</span> && !Bytes.equals(startEndKeys.getSecond()[indexForCallable], HConstants.EMPTY_BYTE_ARRAY)) {<a name="line.728"></a>
-<span class="sourceLineNo">729</span> throw new IOException("The last region info for table "<a name="line.729"></a>
-<span class="sourceLineNo">730</span> + table.getName()<a name="line.730"></a>
-<span class="sourceLineNo">731</span> + " cann't be found in hbase:meta.Please use hbck tool to fix it first.");<a name="line.731"></a>
-<span class="sourceLineNo">732</span> } else if (indexForCallable + 1 < startEndKeys.getFirst().length<a name="line.732"></a>
-<span class="sourceLineNo">733</span> && !(Bytes.compareTo(startEndKeys.getSecond()[indexForCallable],<a name="line.733"></a>
-<span class="sourceLineNo">734</span> startEndKeys.getFirst()[indexForCallable + 1]) == 0)) {<a name="line.734"></a>
-<span class="sourceLineNo">735</span> throw new IOException("The endkey of one region for table "<a name="line.735"></a>
-<span class="sourceLineNo">736</span> + table.getName()<a name="line.736"></a>
-<span class="sourceLineNo">737</span> + " is not equal to the startkey of the next region in hbase:meta."<a name="line.737"></a>
-<span class="sourceLineNo">738</span> + "Please use hbck tool to fix it first.");<a name="line.738"></a>
-<span class="sourceLineNo">739</span> }<a name="line.739"></a>
-<span class="sourceLineNo">740</span><a name="line.740"></a>
-<span class="sourceLineNo">741</span> boolean lastKeyInRange =<a name="line.741"></a>
-<span class="sourceLineNo">742</span> Bytes.compareTo(last, startEndKeys.getSecond()[idx]) < 0 ||<a name="line.742"></a>
-<span class="sourceLineNo">743</span> Bytes.equals(startEndKeys.getSecond()[idx], HConstants.EMPTY_BYTE_ARRAY);<a name="line.743"></a>
-<span class="sourceLineNo">744</span> if (!lastKeyInRange) {<a name="line.744"></a>
-<span class="sourceLineNo">745</span> List<LoadQueueItem> lqis = splitStoreFile(item, table,<a name="line.745"></a>
-<span class="sourceLineNo">746</span> startEndKeys.getFirst()[indexForCallable],<a name="line.746"></a>
-<span class="sourceLineNo">747</span> startEndKeys.getSecond()[indexForCallable]);<a name="line.747"></a>
-<span class="sourceLineNo">748</span> return lqis;<a name="line.748"></a>
-<span class="sourceLineNo">749</span> }<a name="line.749"></a>
-<span class="sourceLineNo">750</span><a name="line.750"></a>
-<span class="sourceLineNo">751</span> // group regions.<a name="line.751"></a>
-<span class="sourceLineNo">752</span> regionGroups.put(ByteBuffer.wrap(startEndKeys.getFirst()[idx]), item);<a name="line.752"></a>
-<span class="sourceLineNo">753</span> return null;<a name="line.753"></a>
-<span class="sourceLineNo">754</span> }<a name="line.754"></a>
-<span class="sourceLineNo">755</span><a name="line.755"></a>
-<span class="sourceLineNo">756</span> /**<a name="line.756"></a>
-<span class="sourceLineNo">757</span> * Attempts to do an atomic load of many hfiles into a region. If it fails,<a name="line.757"></a>
-<span class="sourceLineNo">758</span> * it returns a list of hfiles that need to be retried. If it is successful<a name="line.758"></a>
-<span class="sourceLineNo">759</span> * it will return an empty list.<a name="line.759"></a>
-<span class="sourceLineNo">760</span> *<a name="line.760"></a>
-<span class="sourceLineNo">761</span> * NOTE: To maintain row atomicity guarantees, region server callable should<a name="line.761"></a>
-<span class="sourceLineNo">762</span> * succeed atomically and fails atomically.<a name="line.762"></a>
-<span class="sourceLineNo">763</span> *<a name="line.763"></a>
-<span class="sourceLineNo">764</span> * Protected for testing.<a name="line.764"></a>
-<span class="sourceLineNo">765</span> *<a name="line.765"></a>
-<span class="sourceLineNo">766</span> * @return empty list if success, list of items to retry on recoverable<a name="line.766"></a>
-<span class="sourceLineNo">767</span> * failure<a name="line.767"></a>
-<span class="sourceLineNo">768</span> */<a name="line.768"></a>
-<span class="sourceLineNo">769</span> protected List<LoadQueueItem> tryAtomicRegionLoad(final Connection conn,<a name="line.769"></a>
-<span class="sourceLineNo">770</span> final TableName tableName, final byte[] first, final Collection<LoadQueueItem> lqis)<a name="line.770"></a>
-<span class="sourceLineNo">771</span> throws IOException {<a name="line.771"></a>
-<span class="sourceLineNo">772</span> final List<Pair<byte[], String>> famPaths =<a name="line.772"></a>
-<span class="sourceLineNo">773</span> new ArrayList<Pair<byte[], String>>(lqis.size());<a name="line.773"></a>
-<span class="sourceLineNo">774</span> for (LoadQueueItem lqi : lqis) {<a name="line.774"></a>
-<span class="sourceLineNo">775</span> famPaths.add(Pair.newPair(lqi.family, lqi.hfilePath.toString()));<a name="line.775"></a>
-<span class="sourceLineNo">776</span> }<a name="line.776"></a>
-<span class="sourceLineNo">777</span><a name="line.777"></a>
-<span class="sourceLineNo">778</span> final RegionServerCallable<Boolean> svrCallable =<a name="line.778"></a>
-<span class="sourceLineNo">779</span> new RegionServerCallable<Boolean>(conn, tableName, first) {<a name="line.779"></a>
-<span class="sourceLineNo">780</span> @Override<a name="line.780"></a>
-<span class="sourceLineNo">781</span> public Boolean call(int callTimeout) throws Exception {<a name="line.781"></a>
-<span class="sourceLineNo">782</span> SecureBulkLoadClient secureClient = null;<a name="line.782"></a>
-<span class="sourceLineNo">783</span> boolean success = false;<a name="line.783"></a>
-<span class="sourceLineNo">784</span><a name="line.784"></a>
-<span class="sourceLineNo">785</span> try {<a name="line.785"></a>
-<span class="sourceLineNo">786</span> LOG.debug("Going to connect to server " + getLocation() + " for row "<a name="line.786"></a>
-<span class="sourceLineNo">787</span> + Bytes.toStringBinary(getRow()) + " with hfile group " + famPaths);<a name="line.787"></a>
-<span class="sourceLineNo">788</span> byte[] regionName = getLocation().getRegionInfo().getRegionName();<a name="line.788"></a>
-<span class="sourceLineNo">789</span> if (!isSecureBulkLoadEndpointAvailable()) {<a name="line.789"></a>
-<span class="sourceLineNo">790</span> success = ProtobufUtil.bulkLoadHFile(getStub(), famPaths, regionName, assignSeqIds);<a name="line.790"></a>
-<span class="sourceLineNo">791</span> } else {<a name="line.791"></a>
-<span class="sourceLineNo">792</span> try (Table table = conn.getTable(getTableName())) {<a name="line.792"></a>
-<span class="sourceLineNo">793</span> secureClient = new SecureBulkLoadClient(table);<a name="line.793"></a>
-<span class="sourceLineNo">794</span> success = secureClient.bulkLoadHFiles(famPaths, fsDelegationToken.getUserToken(),<a name="line.794"></a>
-<span class="sourceLineNo">795</span> bulkToken, getLocation().getRegionInfo().getStartKey());<a name="line.795"></a>
-<span class="sourceLineNo">796</span> }<a name="line.796"></a>
-<span class="sourceLineNo">797</span> }<a name="line.797"></a>
-<span class="sourceLineNo">798</span> return success;<a name="line.798"></a>
-<span class="sourceLineNo">799</span> } finally {<a name="line.799"></a>
-<span class="sourceLineNo">800</span> //Best effort copying of files that might not have been imported<a name="line.800"></a>
-<span class="sourceLineNo">801</span> //from the staging directory back to original location<a name="line.801"></a>
-<span class="sourceLineNo">802</span> //in user directory<a name="line.802"></a>
-<span class="sourceLineNo">803</span> if(secureClient != null && !success) {<a name="line.803"></a>
-<span class="sourceLineNo">804</span> FileSystem targetFs = FileSystem.get(getConf());<a name="line.804"></a>
-<span class="sourceLineNo">805</span> // fs is the source filesystem<a name="line.805"></a>
-<span class="sourceLineNo">806</span> if(fs == null) {<a name="line.806"></a>
-<span class="sourceLineNo">807</span> fs = lqis.iterator().next().hfilePath.getFileSystem(getConf());<a name="line.807"></a>
-<span class="sourceLineNo">808</span> }<a name="line.808"></a>
-<span class="sourceLineNo">809</span> // Check to see if the source and target filesystems are the same<a name="line.809"></a>
-<span class="sourceLineNo">810</span> // If they are the same filesystem, we will try move the files back<a name="line.810"></a>
-<span class="sourceLineNo">811</span> // because previously we moved them to the staging directory.<a name="line.811"></a>
-<span class="sourceLineNo">812</span> if (FSHDFSUtils.isSameHdfs(getConf(), fs, targetFs)) {<a name="line.812"></a>
-<span class="sourceLineNo">813</span> for(Pair<byte[], String> el : famPaths) {<a name="line.813"></a>
-<span class="sourceLineNo">814</span> Path hfileStagingPath = null;<a name="line.814"></a>
-<span class="sourceLineNo">815</span> Path hfileOrigPath = new Path(el.getSecond());<a name="line.815"></a>
-<span class="sourceLineNo">816</span> try {<a name="line.816"></a>
-<span class="sourceLineNo">817</span> hfileStagingPath= new Path(secureClient.getStagingPath(bulkToken, el.getFirst()),<a name="line.817"></a>
-<span class="sourceLineNo">818</span> hfileOrigPath.getName());<a name="line.818"></a>
-<span class="sourceLineNo">819</span> if(targetFs.rename(hfileStagingPath, hfileOrigPath)) {<a name="line.819"></a>
-<span class="sourceLineNo">820</span> LOG.debug("Moved back file " + hfileOrigPath + " from " +<a name="line.820"></a>
-<span class="sourceLineNo">821</span> hfileStagingPath);<a name="line.821"></a>
-<span class="sourceLineNo">822</span> } else if(targetFs.exists(hfileStagingPath)){<a name="line.822"></a>
-<span class="sourceLineNo">823</span> LOG.debug("Unable to move back file " + hfileOrigPath + " from " +<a name="line.823"></a>
-<span class="sourceLineNo">824</span> hfileStagingPath);<a name="line.824"></a>
-<span class="sourceLineNo">825</span> }<a name="line.825"></a>
-<span class="sourceLineNo">826</span> } catch(Exception ex) {<a name="line.826"></a>
-<span class="sourceLineNo">827</span> LOG.debug("Unable to move back file " + hfileOrigPath + " from " +<a name="line.827"></a>
-<span class="sourceLineNo">828</span> hfileStagingPath, ex);<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> }<a name="line.833"></a>
-<span class="sourceLineNo">834</span> }<a name="line.834"></a>
-<span class="sourceLineNo">835</span> };<a name="line.835"></a>
-<span class="sourceLineNo">836</span><a name="line.836"></a>
-<span class="sourceLineNo">837</span> try {<a name="line.837"></a>
-<span class="sourceLineNo">838</span> List<LoadQueueItem> toRetry = new ArrayList<LoadQueueItem>();<a name="line.838"></a>
-<span class="sourceLineNo">839</span> Configuration conf = getConf();<a name="line.839"></a>
-<span class="sourceLineNo">840</span> boolean success = RpcRetryingCallerFactory.instantiate(conf,<a name="line.840"></a>
-<span class="sourceLineNo">841</span> null).<Boolean> newCaller()<a name="line.841"></a>
-<span class="sourceLineNo">842</span> .callWithRetries(svrCallable, Integer.MAX_VALUE);<a name="line.842"></a>
-<span class="sourceLineNo">843</span> if (!success) {<a name="line.843"></a>
-<span class="sourceLineNo">844</span> LOG.warn("Attempt to bulk load region containing "<a name="line.844"></a>
-<span class="sourceLineNo">845</span> + Bytes.toStringBinary(first) + " into table "<a name="line.845"></a>
-<span class="sourceLineNo">846</span> + tableName + " with files " + lqis<a name="line.846"></a>
-<span class="sourceLineNo">847</span> + " failed. This is recoverable and they will be retried.");<a name="line.847"></a>
-<span class="sourceLineNo">848</span> toRetry.addAll(lqis); // return lqi's to retry<a name="line.848"></a>
-<span class="sourceLineNo">849</span> }<a name="line.849"></a>
-<span class="sourceLineNo">850</span> // success<a name="line.850"></a>
-<span class="sourceLineNo">851</span> return toRetry;<a name="line.851"></a>
-<span class="sourceLineNo">852</span> } catch (IOException e) {<a name="line.852"></a>
-<span class="sourceLineNo">853</span> LOG.error("Encountered unrecoverable error from region server, additional details: "<a name="line.853"></a>
-<span class="sourceLineNo">854</span> + svrCallable.getExceptionMessageAdditionalDetail(), e);<a name="line.854"></a>
-<span class="sourceLineNo">855</span> throw e;<a name="line.855"></a>
-<span class="sourceLineNo">856</span> }<a name="line.856"></a>
-<span class="sourceLineNo">857</span> }<a name="line.857"></a>
-<span class="sourceLineNo">858</span><a name="line.858"></a>
-<span class="sourceLineNo">859</span> private boolean isSecureBulkLoadEndpointAvailable() {<a name="line.859"></a>
-<span class="sourceLineNo">860</span> String classes = getConf().get(CoprocessorHost.REGION_COPROCESSOR_CONF_KEY, "");<a name="line.860"></a>
-<span class="sourceLineNo">861</span> return classes.contains("org.apache.hadoop.hbase.security.access.SecureBulkLoadEndpoint");<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> * Split a storefile into a top and bottom half, maintaining<a name="line.865"></a>
-<span class="sourceLineNo">866</span> * the metadata, recreating bloom filters, etc.<a name="line.866"></a>
-<span class="sourceLineNo">867</span> */<a name="line.867"></a>
-<span class="sourceLineNo">868</span> static void splitStoreFile(<a name="line.868"></a>
-<span class="sourceLineNo">869</span> Configuration conf, Path inFile,<a name="line.869"></a>
-<span class="sourceLineNo">870</span> HColumnDescriptor familyDesc, byte[] splitKey,<a name="line.870"></a>
-<span class="sourceLineNo">871</span> Path bottomOut, Path topOut) throws IOException<a name="line.871"></a>
-<span class="sourceLineNo">872</span> {<a name="line.872"></a>
-<span class="sourceLineNo">873</span> // Open reader with no block cache, and not in-memory<a name="line.873"></a>
-<span class="sourceLineNo">874</span> Reference topReference = Reference.createTopReference(splitKey);<a name="line.874"></a>
-<span class="sourceLineNo">875</span> Reference bottomReference = Reference.createBottomReference(splitKey);<a name="line.875"></a>
-<span class="sourceLineNo">876</span><a name="line.876"></a>
-<span class="sourceLineNo">877</span> copyHFileHalf(conf, inFile, topOut, topReference, familyDesc);<a name="line.877"></a>
-<span class="sourceLineNo">878</span> copyHFileHalf(conf, inFile, bottomOut, bottomReference, familyDesc);<a name="line.878"></a>
-<span class="sourceLineNo">879</span> }<a name="line.879"></a>
-<span class="sourceLineNo">880</span><a name="line.880"></a>
-<span class="sourceLineNo">881</span> /**<a name="line.881"></a>
-<span class="sourceLineNo">882</span> * Copy half of an HFile into a new HFile.<a name="line.882"></a>
-<span class="sourceLineNo">883</span> */<a name="line.883"></a>
-<span class="sourceLineNo">884</span> private static void copyHFileHalf(<a name="line.884"></a>
-<span class="sourceLineNo">885</span> Configuration conf, Path inFile, Path outFile, Reference reference,<a name="line.885"></a>
-<span class="sourceLineNo">886</span> HColumnDescriptor familyDescriptor)<a name="line.886"></a>
-<span class="sourceLineNo">887</span> throws IOException {<a name="line.887"></a>
-<span class="sourceLineNo">888</span> FileSystem fs = inFile.getFileSystem(conf);<a name="line.888"></a>
-<span class="sourceLineNo">889</span> CacheConfig cacheConf = new CacheConfig(conf);<a name="line.889"></a>
-<span class="sourceLineNo">890</span> HalfStoreFileReader halfReader = null;<a name="line.890"></a>
-<span class="sourceLineNo">891</span> StoreFile.Writer halfWriter = null;<a name="line.891"></a>
-<span class="sourceLineNo">892</span> try {<a name="line.892"></a>
-<span class="sourceLineNo">893</span> halfReader = new HalfStoreFileReader(fs, inFile, cacheConf, reference, conf);<a name="line.893"></a>
-<span class="sourceLineNo">894</span> Map<byte[], byte[]> fileInfo = halfReader.loadFileInfo();<a name="line.894"></a>
-<span class="sourceLineNo">895</span><a name="line.895"></a>
-<span class="sourceLineNo">896</span> int blocksize = familyDescriptor.getBlocksize();<a name="line.896"></a>
-<span class="sourceLineNo">897</span> Algorithm compression = familyDescriptor.getCompressionType();<a name="line.897"></a>
-<span class="sourceLineNo">898</span> BloomType bloomFilterType = familyDescriptor.getBloomFilterType();<a name="line.898"></a>
-<span class="sourceLineNo">899</span> HFileContext hFileContext = new HFileContextBuilder()<a name="line.899"></a>
-<span class="sourceLineNo">900</span> .withCompression(compression)<a name="line.900"></a>
-<span class="sourceLineNo">901</span> .withChecksumType(HStore.getChecksumType(conf))<a name="line.901"></a>
-<span class="sourceLineNo">902</span> .withBytesPerCheckSum(HStore.getBytesPerChecksum(conf))<a name="line.902"></a>
-<span class="sourceLineNo">903</span> .withBlockSize(blocksize)<a name="line.903"></a>
-<span class="sourceLineNo">904</span> .withDataBlockEncoding(familyDescriptor.getDataBlockEncoding())<a name="line.904"></a>
-<span class="sourceLineNo">905</span> .build();<a name="line.905"></a>
-<span class="sourceLineNo">906</span> halfWriter = new StoreFile.WriterBuilder(conf, cacheConf,<a name="line.906"></a>
-<span class="sourceLineNo">907</span> fs)<a name="line.907"></a>
-<span class="sourceLineNo">908</span> .withFilePath(outFile)<a name="line.908"></a>
-<span class="sourceLineNo">909</span> .withBloomType(bloomFilterType)<a name="line.909"></a>
-<span class="sourceLineNo">910</span> .withFileContext(hFileContext)<a name="line.910"></a>
-<span class="sourceLineNo">911</span> .build();<a name="line.911"></a>
-<span class="sourceLineNo">912</span> HFileScanner scanner = halfReader.getScanner(false, false, false);<a name="line.912"></a>
-<span class="sourceLineNo">913</span> scanner.seekTo();<a name="line.913"></a>
-<span class="sourceLineNo">914</span> do {<a name="line.914"></a>
-<span class="sourceLineNo">915</span> halfWriter.append(scanner.getCell());<a name="line.915"></a>
-<span class="sourceLineNo">916</span> } while (scanner.next());<a name="line.916"></a>
-<span class="sourceLineNo">917</span><a name="line.917"></a>
-<span class="sourceLineNo">918</span> for (Map.Entry<byte[],byte[]> entry : fileInfo.entrySet()) {<a name="line.918"></a>
-<span class="sourceLineNo">919</span> if (shouldCopyHFileMetaKey(entry.getKey())) {<a name="line.919"></a>
-<span class="sourceLineNo">920</span> halfWriter.appendFileInfo(entry.getKey(), entry.getValue());<a name="line.920"></a>
-<span class="sourceLineNo">921</span> }<a name="line.921"></a>
-<span class="sourceLineNo">922</span> }<a name="line.922"></a>
-<span class="sourceLineNo">923</span> } finally {<a name="line.923"></a>
-<span class="sourceLineNo">924</span> if (halfWriter != null) halfWriter.close();<a name="line.924"></a>
-<span class="sourceLineNo">925</span> if (halfReader != null) halfReader.close(cacheConf.shouldEvictOnClose());<a name="line.925"></a>
-<span class="sourceLineNo">926</span> }<a name="line.926"></a>
-<span class="sourceLineNo">927</span> }<a name="line.927"></a>
-<span class="sourceLineNo">928</span><a name="line.928"></a>
-<span class="sourceLineNo">929</span> private static boolean shouldCopyHFileMetaKey(byte[] key) {<a name="line.929"></a>
-<span class="sourceLineNo">930</span> return !HFile.isReservedFileInfoKey(key);<a name="line.930"></a>
-<span class="sourceLineNo">931</span> }<a name="line.931"></a>
-<span class="sourceLineNo">932</span><a name="line.932"></a>
-<span class="sourceLineNo">933</span> /*<a name="line.933"></a>
-<span class="sourceLineNo">934</span> * Infers region boundaries for a new table.<a name="line.934"></a>
-<span class="sourceLineNo">935</span> * Parameter:<a name="line.935"></a>
-<span class="sourceLineNo">936</span> * bdryMap is a map between keys to an integer belonging to {+1, -1}<a name="line.936"></a>
-<span class="sourceLineNo">937</span> * If a key is a start key of a file, then it maps to +1<a name="line.937"></a>
-<span class="sourceLineNo">938</span> * If a key is an end key of a file, then it maps to -1<a name="line.938"></a>
-<span class="sourceLineNo">939</span> * Algo:<a name="line.939"></a>
-<span class="sourceLineNo">940</span> * 1) Poll on the keys in order:<a name="line.940"></a>
-<span class="sourceLineNo">941</span> * a) Keep adding the mapped values to these keys (runningSum)<a name="line.941"></a>
-<span class="sourceLineNo">942</span> * b) Each time runningSum reaches 0, add the start Key from when the runningSum had started to<a name="line.942"></a>
-<span class="sourceLineNo">943</span> * a boundary list.<a name="line.943"></a>
-<span class="sourceLineNo">944</span> * 2) Return the boundary list.<a name="line.944"></a>
-<span class="sourceLineNo">945</span> */<a name="line.945"></a>
-<span class="sourceLineNo">946</span> public static byte[][] inferBoundaries(TreeMap<byte[], Integer> bdryMap) {<a name="line.946"></a>
-<span class="sourceLineNo">947</span> ArrayList<byte[]> keysArray = new ArrayList<byte[]>();<a name="line.947"></a>
-<span class="sourceLineNo">948</span> int runningValue = 0;<a name="line.948"></a>
-<span class="sourceLineNo">949</span> byte[] currStartKey = null;<a name="line.949"></a>
-<span class="sourceLineNo">950</span> boolean firstBoundary = true;<a name="line.950"></a>
-<span class="sourceLineNo">951</span><a name="line.951"></a>
-<span class="sourceLineNo">952</span> for (Map.Entry<byte[], Integer> item: bdryMap.entrySet()) {<a name="line.952"></a>
-<span class="sourceLineNo">953</span> if (runningValue == 0) currStartKey = item.getKey();<a name="line.953"></a>
-<span class="sourceLineNo">954</span> runningValue += item.getValue();<a name="line.954"></a>
-<span class="sourceLineNo">955</span> if (runningValue == 0) {<a name="line.955"></a>
-<span class="sourceLineNo">956</span> if (!firstBoundary) keysArray.add(currStartKey);<a name="line.956"></a>
-<span class="sourceLineNo">957</span> firstBoundary = false;<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> return keysArray.toArray(new byte[0][0]);<a name="line.961"></a>
-<span class="sourceLineNo">962</span> }<a name="line.962"></a>
-<span class="sourceLineNo">963</span><a name="line.963"></a>
-<span class="sourceLineNo">964</span> /*<a name="line.964"></a>
-<span class="sourceLineNo">965</span> * If the table is created for the first time, then "completebulkload" reads the files twice.<a name="line.965"></a>
-<span class="sourceLineNo">966</span> * More modifications necessary if we want to avoid doing it.<a name="line.966"></a>
-<span class="sourceLineNo">967</span> */<a name="line.967"></a>
-<span class="sourceLineNo">968</span> private void createTable(TableName tableName, String dirPath, Admin admin) throws Exception {<a name="line.968"></a>
-<span class="sourceLineNo">969</span> final Path hfofDir = new Path(dirPath);<a name="line.969"></a>
-<span class="sourceLineNo">970</span> final FileSystem fs = hfofDir.getFileSystem(getConf());<a name="line.970"></a>
-<span class="sourceLineNo">971</span><a name="line.971"></a>
-<span class="sourceLineNo">972</span> // Add column families<a name="line.972"></a>
-<span class="sourceLineNo">973</span> // Build a set of keys<a name="line.973"></a>
-<span class="sourceLineNo">974</span> final HTableDescriptor htd = new HTableDescriptor(tableName);<a name="line.974"></a>
-<span class="sourceLineNo">975</span> final TreeMap<byte[], Integer> map = new TreeMap<byte[], Integer>(Bytes.BYTES_COMPARATOR);<a name="line.975"></a>
-<span class="sourceLineNo">976</span> visitBulkHFiles(fs, hfofDir, new BulkHFileVisitor<HColumnDescriptor>() {<a name="line.976"></a>
-<span class="sourceLineNo">977</span> @Override<a name="line.977"></a>
-<span class="sourceLineNo">978</span> public HColumnDescriptor bulkFamily(final byte[] familyName) {<a name="line.978"></a>
-<span class="sourceLineNo">979</span> HColumnDescriptor hcd = new HColumnDescriptor(familyName);<a name="line.979"></a>
-<span class="sourceLineNo">980</span> htd.addFamily(hcd);<a name="line.980"></a>
-<span class="sourceLineNo">981</span> return hcd;<a name="line.981"></a>
-<span class="sourceLineNo">982</span> }<a name="line.982"></a>
-<span class="sourceLineNo">983</span> @Override<a name="line.983"></a>
-<span class="sourceLineNo">984</span> public void bulkHFile(final HColumnDescriptor hcd, final FileStatus hfileStatus)<a name="line.984"></a>
-<span class="sourceLineNo">985</span> throws IOException {<a name="line.985"></a>
-<span class="sourceLineNo">986</span> Path hfile = hfileStatus.getPath();<a name="line.986"></a>
-<span class="sourceLineNo">987</span> HFile.Reader reader = HFile.createReader(fs, hfile,<a name="line.987"></a>
-<span class="sourceLineNo">988</span> new CacheConfig(getConf()), getConf());<a name="line.988"></a>
-<span class="sourceLineNo">989</span> try {<a name="line.989"></a>
-<span class="sourceLineNo">990</span> if (hcd.getCompressionType() != reader.getFileContext().getCompression()) {<a name="line.990"></a>
-<span class="sourceLineNo">991</span> hcd.setCompressionType(reader.getFileContext().getCompression());<a name="line.991"></a>
-<span class="sourceLineNo">992</span> LOG.info("Setting compression " + hcd.getCompressionType().name() +<a name="line.992"></a>
-<span class="sourceLineNo">993</span> " for family " + hcd.toString());<a name="line.993"></a>
-<span class="sourceLineNo">994</span> }<a name="line.994"></a>
-<span class="sourceLineNo">995</span> reader.loadFileInfo();<a name="line.995"></a>
-<span class="sourceLineNo">996</span> byte[] first = reader.getFirstRowKey();<a name="line.996"></a>
-<span class="sourceLineNo">997</span> byte[] last = reader.getLastRowKey();<a name="line.997"></a>
-<span class="sourceLineNo">998</span><a name="line.998"></a>
-<span class="sourceLineNo">999</span> LOG.info("Trying to figure out region boundaries hfile=" + hfile +<a name="line.999"></a>
-<span class="sourceLineNo">1000</span> " first=" + Bytes.toStringBinary(first) +<a name="line.1000"></a>
-<span class="sourceLineNo">1001</span> " last=" + Bytes.toStringBinary(last));<a name="line.1001"></a>
-<span class="sourceLineNo">1002</span><a name="line.1002"></a>
-<span class="sourceLineNo">1003</span> // To eventually infer start key-end key boundaries<a name="line.1003"></a>
-<span class="sourceLineNo">1004</span> Integer value = map.containsKey(first)? map.get(first):0;<a name="line.1004"></a>
-<span class="sourceLineNo">1005</span> map.put(first, value+1);<a name="line.1005"></a>
-<span class="sourceLineNo">1006</span><a name="line.1006"></a>
-<span class="sourceLineNo">1007</span> value = map.containsKey(last)? map.get(last):0;<a name="line.1007"></a>
-<span class="sourceLineNo">1008</span> map.put(last, value-1);<a name="line.1008"></a>
-<span class="sourceLineNo">1009</span> } finally {<a name="line.1009"></a>
-<span class="sourceLineNo">1010</span> reader.close();<a name="line.1010"></a>
-<span class="sourceLineNo">1011</span> }<a name="line.1011"></a>
-<span class="sourceLineNo">1012</span> }<a name="line.1012"></a>
-<span class="sourceLineNo">1013</span> });<a name="line.1013"></a>
-<span class="sourceLineNo">1014</span><a name="line.1014"></a>
-<span class="sourceLineNo">1015</span> byte[][] keys = LoadIncrementalHFiles.inferBoundaries(map);<a name="line.1015"></a>
-<span class="sourceLineNo">1016</span> admin.createTable(htd, keys);<a name="line.1016"></a>
-<span class="sourceLineNo">1017</span><a name="line.1017"></a>
-<span class="sourceLineNo">1018</span> LOG.info("Table "+ tableName +" is available!!");<a name="line.1018"></a>
-<span class="sourceLineNo">1019</span> }<a name="line.1019"></a>
-<span class="sourceLineNo">1020</span><a name="line.1020"></a>
-<span class="sourceLineNo">1021</span> @Override<a name="line.1021"></a>
-<span class="sourceLineNo">1022</span> public int run(String[] args) throws Exception {<a name="line.1022"></a>
-<span class="sourceLineNo">1023</span> if (args.length != 2) {<a name="line.1023"></a>
-<span class="sourceLineNo">1024</span> usage();<a name="line.1024"></a>
-<span class="sourceLineNo">1025</span> return -1;<a name="line.1025"></a>
-<span class="sourceLineNo">1026</span> }<a name="line.1026"></a>
-<span class="sourceLineNo">1027</span><a name="line.1027"></a>
-<span class="sourceLineNo">1028</span> initialize();<a name="line.1028"></a>
-<span class="sourceLineNo">1029</span> try (Connection connection = ConnectionFactory.createConnection(getConf());<a name="line.1029"></a>
-<span class="sourceLineNo">1030</span> Admin admin = connection.getAdmin()) {<a name="line.1030"></a>
-<span class="sourceLineNo">1031</span> String dirPath = args[0];<a name="line.1031"></a>
-<span class="sourceLineNo">1032</span> TableName tableName = TableName.valueOf(args[1]);<a name="line.1032"></a>
-<span class="sourceLineNo">1033</span><a name="line.1033"></a>
-<span class="sourceLineNo">1034</span> boolean tableExists = admin.tableExists(tableName);<a name="line.1034"></a>
-<span class="sourceLineNo">1035</span> if (!tableExists) {<a name="line.1035"></a>
-<span class="sourceLineNo">1036</span> if ("yes".equalsIgnoreCase(getConf().get(CREATE_TABLE_CONF_KEY, "yes"))) {<a name="line.1036"></a>
-<span class="sourceLineNo">1037</span> this.createTable(tableName, dirPath, admin);<a name="line.1037"></a>
-<span class="sourceLineNo">1038</span> } else {<a name="line.1038"></a>
-<span class="sourceLineNo">1039</span> String errorMsg = format("Table '%s' does not exist.", tableName);<a name="line.1039"></a>
-<span class="sourceLineNo">1040</span> LOG.error(errorMsg);<a name="line.1040"></a>
-<span class="sourceLineNo">1041</span> throw new TableNotFoundException(errorMsg);<a name="line.1041"></a>
-<span class="sourceLineNo">1042</span> }<a name="line.1042"></a>
-<span class="sourceLineNo">1043</span> }<a name="line.1043"></a>
-<span class="sourceLineNo">1044</span><a name="line.1044"></a>
-<span class="sourceLineNo">1045</span> Path hfofDir = new Path(dirPath);<a name="line.1045"></a>
-<span class="sourceLineNo">1046</span><a name="line.1046"></a>
-<span class="sourceLineNo">1047</span> try (Table table = connection.getTable(tableName);<a name="line.1047"></a>
-<span class="sourceLineNo">1048</span> RegionLocator locator = connection.getRegionLocator(tableName)) {<a name="line.1048"></a>
-<span class="sourceLineNo">1049</span> doBulkLoad(hfofDir, admin, table, locator);<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> return 0;<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> public static void main(String[] args) throws Exception {<a name="line.1056"></a>
-<span class="sourceLineNo">1057</span> Configuration conf = HBaseConfiguration.create();<a name="line.1057"></a>
-<span class="sourceLineNo">1058</span> int ret = ToolRunner.run(conf, new LoadIncrementalHFiles(), args);<a name="line.1058"></a>
-<span class="sourceLineNo">1059</span> System.exit(ret);<a name="line.1059"></a>
-<span class="sourceLineNo">1060</span> }<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> * Called from replication sink, where it manages bulkToken(staging directory) by itself. This is<a name="line.1063"></a>
-<span class="sourceLineNo">1064</span> * used only when {@link SecureBulkLoadEndpoint} is configured in hbase.coprocessor.region.classes<a name="line.1064"></a>
-<span class="sourceLineNo">1065</span> * property. This directory is used as a temporary directory where all files are initially<a name="line.1065"></a>
-<span class="sourceLineNo">1066</span> * copied/moved from user given directory, set all the required file permissions and then from<a name="line.1066"></a>
-<span class="sourceLineNo">1067</span> * their it is finally loaded into a table. This should be set only when, one would like to manage<a name="line.1067"></a>
-<span class="sourceLineNo">1068</span> * the staging directory by itself. Otherwise this tool will handle this by itself.<a name="line.1068"></a>
-<span class="sourceLineNo">1069</span> * @param stagingDir staging directory path<a name="line.1069"></a>
-<span class="sourceLineNo">1070</span> */<a name="line.1070"></a>
-<span class="sourceLineNo">1071</span> public void setBulkToken(String stagingDir) {<a name="line.1071"></a>
-<span class="sourceLineNo">1072</span> this.bulkToken = stagingDir;<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">428</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.428"></a>
+<span class="sourceLineNo">429</span> * @throws IOException If any I/O or network error occurred<a name="line.429"></a>
+<span class="sourceLineNo">430</span> */<a name="line.430"></a>
+<span class="sourceLineNo">431</span> public void prepareHFileQueue(Path hfilesDir, Table table, Deque<LoadQueueItem> queue,<a name="line.431"></a>
+<span class="sourceLineNo">432</span> boolean validateHFile) throws IOException {<a name="line.432"></a>
+<span class="sourceLineNo">433</span> discoverLoadQueue(queue, hfilesDir, validateHFile);<a name="line.433"></a>
+<span class="sourceLineNo">434</span> validateFamiliesInHFiles(table, queue);<a name="line.434"></a>
+<span class="sourceLineNo">435</span> }<a name="line.435"></a>
+<span class="sourceLineNo">436</span><a name="line.436"></a>
+<span class="sourceLineNo">437</span> // Initialize a thread pool<a name="line.437"></a>
+<span class="sourceLineNo">438</span> private ExecutorService createExecutorService() {<a name="line.438"></a>
+<span class="sourceLineNo">439</span> ThreadFactoryBuilder builder = new ThreadFactoryBuilder();<a name="line.439"></a>
+<span class="sourceLineNo">440</span> builder.setNameFormat("LoadIncrementalHFiles-%1$d");<a name="line.440"></a>
+<span class="sourceLineNo">441</span> ExecutorService pool = new ThreadPoolExecutor(nrThreads, nrThreads, 60, TimeUnit.SECONDS,<a name="line.441"></a>
+<span class="sourceLineNo">442</span> new LinkedBlockingQueue<Runnable>(), builder.build());<a name="line.442"></a>
+<span class="sourceLineNo">443</span> ((ThreadPoolExecutor) pool).allowCoreThreadTimeOut(true);<a name="line.443"></a>
+<span class="sourceLineNo">444</span> return pool;<a name="line.444"></a>
+<span class="sourceLineNo">445</span> }<a name="line.445"></a>
+<span class="sourceLineNo">446</span><a name="line.446"></a>
+<span class="sourceLineNo">447</span> /**<a name="line.447"></a>
+<span class="sourceLineNo">448</span> * Checks whether there is any invalid family name in HFiles to be bulk loaded.<a name="line.448"></a>
+<span class="sourceLineNo">449</span> */<a name="line.449"></a>
+<span class="sourceLineNo">450</span> private void validateFamiliesInHFiles(Table table, Deque<LoadQueueItem> queue)<a name="line.450"></a>
+<span class="sourceLineNo">451</span> throws IOException {<a name="line.451"></a>
+<span class="sourceLineNo">452</span> Collection<HColumnDescriptor> families = table.getTableDescriptor().getFamilies();<a name="line.452"></a>
+<span class="sourceLineNo">453</span> List<String> familyNames = new ArrayList<String>(families.size());<a name="line.453"></a>
+<span class="sourceLineNo">454</span> for (HColumnDescriptor family : families) {<a name="line.454"></a>
+<span class="sourceLineNo">455</span> familyNames.add(family.getNameAsString());<a name="line.455"></a>
+<span class="sourceLineNo">456</span> }<a name="line.456"></a>
+<span class="sourceLineNo">457</span> List<String> unmatchedFamilies = new ArrayList<String>();<a name="line.457"></a>
+<span class="sourceLineNo">458</span> Iterator<LoadQueueItem> queueIter = queue.iterator();<a name="line.458"></a>
+<span class="sourceLineNo">459</span> while (queueIter.hasNext()) {<a name="line.459"></a>
+<span class="sourceLineNo">460</span> LoadQueueItem lqi = queueIter.next();<a name="line.460"></a>
+<span class="sourceLineNo">461</span> String familyNameInHFile = Bytes.toString(lqi.family);<a name="line.461"></a>
+<span class="sourceLineNo">462</span> if (!familyNames.contains(familyNameInHFile)) {<a name="line.462"></a>
+<span class="sourceLineNo">463</span> unmatchedFamilies.add(familyNameInHFile);<a name="line.463"></a>
+<span class="sourceLineNo">464</span> }<a name="line.464"></a>
+<span class="sourceLineNo">465</span> }<a name="line.465"></a>
+<span class="sourceLineNo">466</span> if (unmatchedFamilies.size() > 0) {<a name="line.466"></a>
+<span class="sourceLineNo">467</span> String msg =<a name="line.467"></a>
+<span class="sourceLineNo">468</span> "Unmatched family names found: unmatched family names in HFiles to be bulkloaded: "<a name="line.468"></a>
+<span class="sourceLineNo">469</span> + unmatchedFamilies + "; valid family names of table " + table.getName() + " are: "<a name="line.469"></a>
+<span class="sourceLineNo">470</span> + familyNames;<a name="line.470"></a>
+<span class="sourceLineNo">471</span> LOG.error(msg);<a name="line.471"></a>
+<span class="sourceLineNo">472</span> throw new IOException(msg);<a name="line.472"></a>
+<span class="sourceLineNo">473</span> }<a name="line.473"></a>
+<span class="sourceLineNo">474</span> }<a name="line.474"></a>
+<span class="sourceLineNo">475</span><a name="line.475"></a>
+<span class="sourceLineNo">476</span> /**<a name="line.476"></a>
+<span class="sourceLineNo">477</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.477"></a>
+<span class="sourceLineNo">478</span> * 1. {@link LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)} 2.<a name="line.478"></a>
+<span class="sourceLineNo">479</span> * {@link<a name="line.479"></a>
+<span class="sourceLineNo">480</span> * LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)}<a name="line.480"></a>
+<span class="sourceLineNo">481</span> * @param table Table to which these hfiles should be loaded to<a name="line.481"></a>
+<span class="sourceLineNo">482</span> * @param conn Connection to use<a name="line.482"></a>
+<span class="sourceLineNo">483</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.483"></a>
+<span class="sourceLineNo">484</span> * @param startEndKeys starting and ending row keys of the region<a name="line.484"></a>
+<span class="sourceLineNo">485</span> */<a name="line.485"></a>
+<span class="sourceLineNo">486</span> public void loadHFileQueue(final Table table, final Connection conn, Deque<LoadQueueItem> queue,<a name="line.486"></a>
+<span class="sourceLineNo">487</span> Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.487"></a>
+<span class="sourceLineNo">488</span> ExecutorService pool = null;<a name="line.488"></a>
+<span class="sourceLineNo">489</span> try {<a name="line.489"></a>
+<span class="sourceLineNo">490</span> pool = createExecutorService();<a name="line.490"></a>
+<span class="sourceLineNo">491</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups =<a name="line.491"></a>
+<span class="sourceLineNo">492</span> groupOrSplitPhase(table, pool, queue, startEndKeys);<a name="line.492"></a>
+<span class="sourceLineNo">493</span> bulkLoadPhase(table, conn, pool, queue, regionGroups);<a name="line.493"></a>
+<span class="sourceLineNo">494</span> } finally {<a name="line.494"></a>
+<span class="sourceLineNo">495</span> if (pool != null) {<a name="line.495"></a>
+<span class="sourceLineNo">496</span> pool.shutdown();<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> }<a name="line.499"></a>
+<span class="sourceLineNo">500</span><a name="line.500"></a>
+<span class="sourceLineNo">501</span> /**<a name="line.501"></a>
+<span class="sourceLineNo">502</span> * This takes the LQI's grouped by likely regions and attempts to bulk load<a name="line.502"></a>
+<span class="sourceLineNo">503</span> * them. Any failures are re-queued for another pass with the<a name="line.503"></a>
+<span class="sourceLineNo">504</span> * groupOrSplitPhase.<a name="line.504"></a>
+<span class="sourceLineNo">505</span> */<a name="line.505"></a>
+<span class="sourceLineNo">506</span> protected void bulkLoadPhase(final Table table, final Connection conn,<a name="line.506"></a>
+<span class="sourceLineNo">507</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.507"></a>
+<span class="sourceLineNo">508</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) throws IOException {<a name="line.508"></a>
+<span class="sourceLineNo">509</span> // atomically bulk load the groups.<a name="line.509"></a>
+<span class="sourceLineNo">510</span> Set<Future<List<LoadQueueItem>>> loadingFutures = new HashSet<Future<List<LoadQueueItem>>>();<a name="line.510"></a>
+<span class="sourceLineNo">511</span> for (Entry<ByteBuffer, ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()){<a name="line.511"></a>
+<span class="sourceLineNo">512</span> final byte[] first = e.getKey().array();<a name="line.512"></a>
+<span class="sourceLineNo">513</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.513"></a>
+<span class="sourceLineNo">514</span><a name="line.514"></a>
+<span class="sourceLineNo">515</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.515"></a>
+<span class="sourceLineNo">516</span> @Override<a name="line.516"></a>
+<span class="sourceLineNo">517</span> public List<LoadQueueItem> call() throws Exception {<a name="line.517"></a>
+<span class="sourceLineNo">518</span> List<LoadQueueItem> toRetry =<a name="line.518"></a>
+<span class="sourceLineNo">519</span> tryAtomicRegionLoad(conn, table.getName(), first, lqis);<a name="line.519"></a>
+<span class="sourceLineNo">520</span> return toRetry;<a name="line.520"></a>
+<span class="sourceLineNo">521</span> }<a name="line.521"></a>
+<span class="sourceLineNo">522</span> };<a name="line.522"></a>
+<span class="sourceLineNo">523</span> loadingFutures.add(pool.submit(call));<a name="line.523"></a>
+<span class="sourceLineNo">524</span> }<a name="line.524"></a>
+<span class="sourceLineNo">525</span><a name="line.525"></a>
+<span class="sourceLineNo">526</span> // get all the results.<a name="line.526"></a>
+<span class="sourceLineNo">527</span> for (Future<List<LoadQueueItem>> future : loadingFutures) {<a name="line.527"></a>
+<span class="sourceLineNo">528</span> try {<a name="line.528"></a>
+<span class="sourceLineNo">529</span> List<LoadQueueItem> toRetry = future.get();<a name="line.529"></a>
+<span class="sourceLineNo">530</span><a name="line.530"></a>
+<span class="sourceLineNo">531</span> // LQIs that are requeued to be regrouped.<a name="line.531"></a>
+<span class="sourceLineNo">532</span> queue.addAll(toRetry);<a name="line.532"></a>
+<span class="sourceLineNo">533</span><a name="line.533"></a>
+<span class="sourceLineNo">534</span> } catch (ExecutionException e1) {<a name="line.534"></a>
+<span class="sourceLineNo">535</span> Throwable t = e1.getCause();<a name="line.535"></a>
+<span class="sourceLineNo">536</span> if (t instanceof IOException) {<a name="line.536"></a>
+<span class="sourceLineNo">537</span> // At this point something unrecoverable has happened.<a name="line.537"></a>
+<span class="sourceLineNo">538</span> // TODO Implement bulk load recovery<a name="line.538"></a>
+<span class="sourceLineNo">539</span> throw new IOException("BulkLoad encountered an unrecoverable problem", t);<a name="line.539"></a>
+<span class="sourceLineNo">540</span> }<a name="line.540"></a>
+<span class="sourceLineNo">541</span> LOG.error("Unexpected execution exception during bulk load", e1);<a name="line.541"></a>
+<span class="sourceLineNo">542</span> throw new IllegalStateException(t);<a name="line.542"></a>
+<span class="sourceLineNo">543</span> } catch (InterruptedException e1) {<a name="line.543"></a>
+<span class="sourceLineNo">544</span> LOG.error("Unexpected interrupted exception during bulk load", e1);<a name="line.544"></a>
+<span class="sourceLineNo">545</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<a name="line.545"></a>
+<span class="sourceLineNo">546</span> }<a name="line.546"></a>
+<span class="sourceLineNo">547</span> }<a name="line.547"></a>
+<span class="sourceLineNo">548</span> }<a name="line.548"></a>
+<span class="sourceLineNo">549</span><a name="line.549"></a>
+<span class="sourceLineNo">550</span> private boolean checkHFilesCountPerRegionPerFamily(<a name="line.550"></a>
+<span class="sourceLineNo">551</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) {<a name="line.551"></a>
+<span class="sourceLineNo">552</span> for (Entry<ByteBuffer,<a name="line.552"></a>
+<span class="sourceLineNo">553</span> ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()) {<a name="line.553"></a>
+<span class="sourceLineNo">554</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.554"></a>
+<span class="sourceLineNo">555</span> HashMap<byte[], MutableInt> filesMap = new HashMap<byte[], MutableInt>();<a name="line.555"></a>
+<span class="sourceLineNo">556</span> for (LoadQueueItem lqi: lqis) {<a name="line.556"></a>
+<span class="sourceLineNo">557</span> MutableInt count = filesMap.get(lqi.family);<a name="line.557"></a>
+<span class="sourceLineNo">558</span> if (count == null) {<a name="line.558"></a>
+<span class="sourceLineNo">559</span> count = new MutableInt();<a name="line.559"></a>
+<span class="sourceLineNo">560</span> filesMap.put(lqi.family, count);<a name="line.560"></a>
+<span class="sourceLineNo">561</span> }<a name="line.561"></a>
+<span class="sourceLineNo">562</span> count.increment();<a name="line.562"></a>
+<span class="sourceLineNo">563</span> if (count.intValue() > maxFilesPerRegionPerFamily) {<a name="line.563"></a>
+<span class="sourceLineNo">564</span> LOG.error("Trying to load more than " + maxFilesPerRegionPerFamily<a name="line.564"></a>
+<span class="sourceLineNo">565</span> + " hfiles to family " + Bytes.toStringBinary(lqi.family)<a name="line.565"></a>
+<span class="sourceLineNo">566</span> + " of region with start key "<a name="line.566"></a>
+<span class="sourceLineNo">567</span> + Bytes.toStringBinary(e.getKey()));<a name="line.567"></a>
+<span class="sourceLineNo">568</span> return false;<a name="line.568"></a>
+<span class="sourceLineNo">569</span> }<a name="line.569"></a>
+<span class="sourceLineNo">570</span> }<a name="line.570"></a>
+<span class="sourceLineNo">571</span> }<a name="line.571"></a>
+<span class="sourceLineNo">572</span> return true;<a name="line.572"></a>
+<span class="sourceLineNo">573</span> }<a name="line.573"></a>
+<span class="sourceLineNo">574</span><a name="line.574"></a>
+<span class="sourceLineNo">575</span> /**<a name="line.575"></a>
+<span class="sourceLineNo">576</span> * @return A map that groups LQI by likely bulk load region targets.<a name="line.576"></a>
+<span class="sourceLineNo">577</span> */<a name="line.577"></a>
+<span class="sourceLineNo">578</span> private Multimap<ByteBuffer, LoadQueueItem> groupOrSplitPhase(final Table table,<a name="line.578"></a>
+<span class="sourceLineNo">579</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.579"></a>
+<span class="sourceLineNo">580</span> final Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.580"></a>
+<span class="sourceLineNo">581</span> // <region start key, LQI> need synchronized only within this scope of this<a name="line.581"></a>
+<span class="sourceLineNo">582</span> // phase because of the puts that happen in futures.<a name="line.582"></a>
+<span class="sourceLineNo">583</span> Multimap<ByteBuffer, LoadQueueItem> rgs = HashMultimap.create();<a name="line.583"></a>
+<span class="sourceLineNo">584</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups = Multimaps.synchronizedMultimap(rgs);<a name="line.584"></a>
+<span class="sourceLineNo">585</span><a name="line.585"></a>
+<span class="sourceLineNo">586</span> // drain LQIs and figure out bulk load groups<a name="line.586"></a>
+<span class="sourceLineNo">587</span> Set<Future<List<LoadQueueItem>>> splittingFutures = new HashSet<Future<List<LoadQueueItem>>>();<a name="line.587"></a>
+<span class="sourceLineNo">588</span> while (!queue.isEmpty()) {<a name="line.588"></a>
+<span class="sourceLineNo">589</span> final LoadQueueItem item = queue.remove();<a name="line.589"></a>
+<span class="sourceLineNo">590</span><a name="line.590"></a>
+<span class="sourceLineNo">591</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.591"></a>
+<span class="sourceLineNo">592</span> @Override<a name="line.592"></a>
+<span class="sourceLineNo">593</span> public List<LoadQueueItem> call() throws Exception {<a name="line.593"></a>
+<span class="sourceLineNo">594</span> List<LoadQueueItem> splits = groupOrSplit(regionGroups, item, table, startEndKeys);<a name="line.594"></a>
+<span class="sourceLineNo">595</span> return splits;<a name="line.595"></a>
+<span class="sourceLineNo">596</span> }<a name="line.596"></a>
+<span class="sourceLineNo">597</span> };<a name="line.597"></a>
+<span class="sourceLineNo">598</span> splittingFutures.add(pool.submit(call));<a name="line.598"></a>
+<span class="sourceLineNo">599</span> }<a name="line.599"></a>
+<span class="sourceLineNo">600</span> // get all the results. All grouping and splitting must finish before<a name="line.600"></a>
+<span class="sourceLineNo">601</span> // we can attempt the atomic loads.<a name="line.601"></a>
+<span class="sourceLineNo">602</span> for (Future<List<LoadQueueItem>> lqis : splittingFutures) {<a name="line.602"></a>
+<span class="sourceLineNo">603</span> try {<a name="line.603"></a>
+<span class="sourceLineNo">604</span> List<LoadQueueItem> splits = lqis.get();<a name="line.604"></a>
+<span class="sourceLineNo">605</span> if (splits != null) {<a name="line.605"></a>
+<span class="sourceLineNo">606</span> queue.addAll(splits);<a name="line.606"></a>
+<span class="sourceLineNo">607</span> }<a name="line.607"></a>
+<span class="sourceLineNo">608</span> } catch (ExecutionException e1) {<a name="line.608"></a>
+<span class="sourceLineNo">609</span> Throwable t = e1.getCause();<a name="line.609"></a>
+<span class="sourceLineNo">610</span> if (t instanceof IOException) {<a name="line.610"></a>
+<span class="sourceLineNo">611</span> LOG.error("IOException during splitting", e1);<a name="line.611"></a>
+<span class="sourceLineNo">612</span> throw (IOException)t; // would have been thrown if not parallelized,<a name="line.612"></a>
+<span class="sourceLineNo">613</span> }<a name="line.613"></a>
+<span class="sourceLineNo">614</span> LOG.error("Unexpected execution exception during splitting", e1);<a name="line.614"></a>
+<span class="sourceLineNo">615</span> throw new IllegalStateException(t);<a name="line.615"></a>
+<span class="sourceLineNo">616</span> } catch (InterruptedException e1) {<a name="line.616"></a>
+<span class="sourceLineNo">617</span> LOG.error("Unexpected interrupted exception during splitting", e1);<a name="line.617"></a>
+<span class="sourceLineNo">618</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<a name="line.618"></a>
+<span class="sourceLineNo">619</span> }<a name="line.619"></a>
+<span class="sourceLineNo">620</span> }<a name="line.620"></a>
+<span class="sourceLineNo">621</span> return regionGroups;<a name="line.621"></a>
+<span class="sourceLineNo">622</span> }<a name="line.622"></a>
+<span class="sourceLineNo">623</span><a name="line.623"></a>
+<span class="sourceLineNo">624</span> // unique file name for the table<a name="line.624"></a>
+<span class="sourceLineNo">625</span> private String getUniqueName() {<a name="line.625"></a>
+<span class="sourceLineNo">626</span> return UUID.randomUUID().toString().replaceAll("-", "");<a name="line.626"></a>
+<span class="sourceLineNo">627</span> }<a name="line.627"></a>
+<span class="sourceLineNo">628</span><a name="line.628"></a>
+<span class="sourceLineNo">629</span> protected List<LoadQueueItem> splitStoreFile(final LoadQueueItem item,<a name="line.629"></a>
+<span class="sourceLineNo">630</span> final Table table, byte[] startKey,<a name="line.630"></a>
+<span class="sourceLineNo">631</span> byte[] splitKey) throws IOException {<a name="line.631"></a>
+<span class="sourceLineNo">632</span> final Path hfilePath = item.hfilePath;<a name="line.632"></a>
+<span class="sourceLineNo">633</span><a name="line.633"></a>
+<span class="sourceLineNo">634</span> // We use a '_' prefix which is ignored when walking directory trees<a name="line.634"></a>
+<span class="sourceLineNo">635</span> // above.<a name="line.635"></a>
+<span class="sourceLineNo">636</span> final String TMP_DIR
<TRUNCATED>