http://git-wip-us.apache.org/repos/asf/hbase-site/blob/f07ee53f/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html index 07f6210..6d25806 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/LoadIncrementalHFiles.html @@ -370,881 +370,904 @@ <span class="sourceLineNo">362</span> * @param regionLocator region locator<a name="line.362"></a> <span class="sourceLineNo">363</span> * @param silence true to ignore unmatched column families<a name="line.363"></a> <span class="sourceLineNo">364</span> * @param copyFile always copy hfiles if true<a name="line.364"></a> -<span class="sourceLineNo">365</span> * @throws TableNotFoundException if table does not yet exist<a name="line.365"></a> -<span class="sourceLineNo">366</span> */<a name="line.366"></a> -<span class="sourceLineNo">367</span> public void doBulkLoad(Map<byte[], List<Path>> map, final Admin admin, Table table,<a name="line.367"></a> -<span class="sourceLineNo">368</span> RegionLocator regionLocator, boolean silence, boolean copyFile)<a name="line.368"></a> -<span class="sourceLineNo">369</span> throws TableNotFoundException, IOException {<a name="line.369"></a> -<span class="sourceLineNo">370</span> if (!admin.isTableAvailable(regionLocator.getName())) {<a name="line.370"></a> -<span class="sourceLineNo">371</span> throw new TableNotFoundException("Table " + table.getName() + " is not currently available.");<a name="line.371"></a> -<span class="sourceLineNo">372</span> }<a name="line.372"></a> -<span class="sourceLineNo">373</span> // LQI queue does not need to be threadsafe -- all operations on this queue<a name="line.373"></a> -<span class="sourceLineNo">374</span> // happen in this thread<a name="line.374"></a> -<span class="sourceLineNo">375</span> Deque<LoadQueueItem> queue = new LinkedList<>();<a name="line.375"></a> -<span class="sourceLineNo">376</span> ExecutorService pool = null;<a name="line.376"></a> -<span class="sourceLineNo">377</span> SecureBulkLoadClient secureClient = null;<a name="line.377"></a> -<span class="sourceLineNo">378</span> try {<a name="line.378"></a> -<span class="sourceLineNo">379</span> prepareHFileQueue(map, table, queue, silence);<a name="line.379"></a> -<span class="sourceLineNo">380</span> if (queue.isEmpty()) {<a name="line.380"></a> -<span class="sourceLineNo">381</span> LOG.warn("Bulk load operation did not get any files to load");<a name="line.381"></a> -<span class="sourceLineNo">382</span> return;<a name="line.382"></a> -<span class="sourceLineNo">383</span> }<a name="line.383"></a> -<span class="sourceLineNo">384</span> pool = createExecutorService();<a name="line.384"></a> -<span class="sourceLineNo">385</span> secureClient = new SecureBulkLoadClient(table.getConfiguration(), table);<a name="line.385"></a> -<span class="sourceLineNo">386</span> for (Map.Entry<byte[], List<Path>> entry : map.entrySet()) {<a name="line.386"></a> -<span class="sourceLineNo">387</span> for (Path p : entry.getValue()) {<a name="line.387"></a> -<span class="sourceLineNo">388</span> fs = p.getFileSystem(table.getConfiguration());<a name="line.388"></a> -<span class="sourceLineNo">389</span> break;<a name="line.389"></a> -<span class="sourceLineNo">390</span> }<a name="line.390"></a> -<span class="sourceLineNo">391</span> }<a name="line.391"></a> -<span class="sourceLineNo">392</span> performBulkLoad(admin, table, regionLocator, queue, pool, secureClient, copyFile);<a name="line.392"></a> -<span class="sourceLineNo">393</span> } finally {<a name="line.393"></a> -<span class="sourceLineNo">394</span> cleanup(admin, queue, pool, secureClient);<a name="line.394"></a> -<span class="sourceLineNo">395</span> }<a name="line.395"></a> -<span class="sourceLineNo">396</span> }<a name="line.396"></a> -<span class="sourceLineNo">397</span><a name="line.397"></a> -<span class="sourceLineNo">398</span> /**<a name="line.398"></a> -<span class="sourceLineNo">399</span> * Perform a bulk load of the given directory into the given<a name="line.399"></a> -<span class="sourceLineNo">400</span> * pre-existing table. This method is not threadsafe.<a name="line.400"></a> -<span class="sourceLineNo">401</span> *<a name="line.401"></a> -<span class="sourceLineNo">402</span> * @param hfofDir the directory that was provided as the output path<a name="line.402"></a> -<span class="sourceLineNo">403</span> * of a job using HFileOutputFormat<a name="line.403"></a> -<span class="sourceLineNo">404</span> * @param admin the Admin<a name="line.404"></a> -<span class="sourceLineNo">405</span> * @param table the table to load into<a name="line.405"></a> -<span class="sourceLineNo">406</span> * @param regionLocator region locator<a name="line.406"></a> -<span class="sourceLineNo">407</span> * @param silence true to ignore unmatched column families<a name="line.407"></a> -<span class="sourceLineNo">408</span> * @param copyFile always copy hfiles if true<a name="line.408"></a> -<span class="sourceLineNo">409</span> * @throws TableNotFoundException if table does not yet exist<a name="line.409"></a> -<span class="sourceLineNo">410</span> */<a name="line.410"></a> -<span class="sourceLineNo">411</span> public void doBulkLoad(Path hfofDir, final Admin admin, Table table,<a name="line.411"></a> -<span class="sourceLineNo">412</span> RegionLocator regionLocator, boolean silence, boolean copyFile)<a name="line.412"></a> -<span class="sourceLineNo">413</span> throws TableNotFoundException, IOException {<a name="line.413"></a> -<span class="sourceLineNo">414</span> if (!admin.isTableAvailable(regionLocator.getName())) {<a name="line.414"></a> -<span class="sourceLineNo">415</span> throw new TableNotFoundException("Table " + table.getName() + " is not currently available.");<a name="line.415"></a> -<span class="sourceLineNo">416</span> }<a name="line.416"></a> -<span class="sourceLineNo">417</span><a name="line.417"></a> -<span class="sourceLineNo">418</span> /*<a name="line.418"></a> -<span class="sourceLineNo">419</span> * Checking hfile format is a time-consuming operation, we should have an option to skip<a name="line.419"></a> -<span class="sourceLineNo">420</span> * this step when bulkloading millions of HFiles. See HBASE-13985.<a name="line.420"></a> -<span class="sourceLineNo">421</span> */<a name="line.421"></a> -<span class="sourceLineNo">422</span> boolean validateHFile = getConf().getBoolean("hbase.loadincremental.validate.hfile", true);<a name="line.422"></a> -<span class="sourceLineNo">423</span> if (!validateHFile) {<a name="line.423"></a> -<span class="sourceLineNo">424</span> LOG.warn("You are skipping HFiles validation, it might cause some data loss if files " +<a name="line.424"></a> -<span class="sourceLineNo">425</span> "are not correct. If you fail to read data from your table after using this " +<a name="line.425"></a> -<span class="sourceLineNo">426</span> "option, consider removing the files and bulkload again without this option. " +<a name="line.426"></a> -<span class="sourceLineNo">427</span> "See HBASE-13985");<a name="line.427"></a> -<span class="sourceLineNo">428</span> }<a name="line.428"></a> -<span class="sourceLineNo">429</span> // LQI queue does not need to be threadsafe -- all operations on this queue<a name="line.429"></a> -<span class="sourceLineNo">430</span> // happen in this thread<a name="line.430"></a> -<span class="sourceLineNo">431</span> Deque<LoadQueueItem> queue = new LinkedList<>();<a name="line.431"></a> -<span class="sourceLineNo">432</span> ExecutorService pool = null;<a name="line.432"></a> -<span class="sourceLineNo">433</span> SecureBulkLoadClient secureClient = null;<a name="line.433"></a> -<span class="sourceLineNo">434</span> try {<a name="line.434"></a> -<span class="sourceLineNo">435</span> prepareHFileQueue(hfofDir, table, queue, validateHFile, silence);<a name="line.435"></a> -<span class="sourceLineNo">436</span><a name="line.436"></a> -<span class="sourceLineNo">437</span> if (queue.isEmpty()) {<a name="line.437"></a> -<span class="sourceLineNo">438</span> LOG.warn("Bulk load operation did not find any files to load in " +<a name="line.438"></a> -<span class="sourceLineNo">439</span> "directory " + hfofDir != null ? hfofDir.toUri() : "" + ". Does it contain files in " +<a name="line.439"></a> -<span class="sourceLineNo">440</span> "subdirectories that correspond to column family names?");<a name="line.440"></a> -<span class="sourceLineNo">441</span> return;<a name="line.441"></a> -<span class="sourceLineNo">442</span> }<a name="line.442"></a> -<span class="sourceLineNo">443</span> pool = createExecutorService();<a name="line.443"></a> -<span class="sourceLineNo">444</span> secureClient = new SecureBulkLoadClient(table.getConfiguration(), table);<a name="line.444"></a> -<span class="sourceLineNo">445</span> performBulkLoad(admin, table, regionLocator, queue, pool, secureClient, copyFile);<a name="line.445"></a> -<span class="sourceLineNo">446</span> } finally {<a name="line.446"></a> -<span class="sourceLineNo">447</span> cleanup(admin, queue, pool, secureClient);<a name="line.447"></a> -<span class="sourceLineNo">448</span> }<a name="line.448"></a> -<span class="sourceLineNo">449</span> }<a name="line.449"></a> -<span class="sourceLineNo">450</span><a name="line.450"></a> -<span class="sourceLineNo">451</span> void performBulkLoad(final Admin admin, Table table, RegionLocator regionLocator,<a name="line.451"></a> -<span class="sourceLineNo">452</span> Deque<LoadQueueItem> queue, ExecutorService pool,<a name="line.452"></a> -<span class="sourceLineNo">453</span> SecureBulkLoadClient secureClient, boolean copyFile) throws IOException {<a name="line.453"></a> -<span class="sourceLineNo">454</span> int count = 0;<a name="line.454"></a> -<span class="sourceLineNo">455</span><a name="line.455"></a> -<span class="sourceLineNo">456</span> if(isSecureBulkLoadEndpointAvailable()) {<a name="line.456"></a> -<span class="sourceLineNo">457</span> LOG.warn("SecureBulkLoadEndpoint is deprecated. It will be removed in future releases.");<a name="line.457"></a> -<span class="sourceLineNo">458</span> LOG.warn("Secure bulk load has been integrated into HBase core.");<a name="line.458"></a> -<span class="sourceLineNo">459</span> }<a name="line.459"></a> -<span class="sourceLineNo">460</span><a name="line.460"></a> -<span class="sourceLineNo">461</span> //If using secure bulk load, get source delegation token, and<a name="line.461"></a> -<span class="sourceLineNo">462</span> //prepare staging directory and token<a name="line.462"></a> -<span class="sourceLineNo">463</span> // fs is the source filesystem<a name="line.463"></a> -<span class="sourceLineNo">464</span> fsDelegationToken.acquireDelegationToken(fs);<a name="line.464"></a> -<span class="sourceLineNo">465</span> bulkToken = secureClient.prepareBulkLoad(admin.getConnection());<a name="line.465"></a> -<span class="sourceLineNo">466</span><a name="line.466"></a> -<span class="sourceLineNo">467</span> // Assumes that region splits can happen while this occurs.<a name="line.467"></a> -<span class="sourceLineNo">468</span> while (!queue.isEmpty()) {<a name="line.468"></a> -<span class="sourceLineNo">469</span> // need to reload split keys each iteration.<a name="line.469"></a> -<span class="sourceLineNo">470</span> final Pair<byte[][], byte[][]> startEndKeys = regionLocator.getStartEndKeys();<a name="line.470"></a> -<span class="sourceLineNo">471</span> if (count != 0) {<a name="line.471"></a> -<span class="sourceLineNo">472</span> LOG.info("Split occured while grouping HFiles, retry attempt " +<a name="line.472"></a> -<span class="sourceLineNo">473</span> + count + " with " + queue.size() + " files remaining to group or split");<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> int maxRetries = getConf().getInt(HConstants.BULKLOAD_MAX_RETRIES_NUMBER, 10);<a name="line.476"></a> -<span class="sourceLineNo">477</span> maxRetries = Math.max(maxRetries, startEndKeys.getFirst().length + 1);<a name="line.477"></a> -<span class="sourceLineNo">478</span> if (maxRetries != 0 && count >= maxRetries) {<a name="line.478"></a> -<span class="sourceLineNo">479</span> throw new IOException("Retry attempted " + count +<a name="line.479"></a> -<span class="sourceLineNo">480</span> " times without completing, bailing out");<a name="line.480"></a> -<span class="sourceLineNo">481</span> }<a name="line.481"></a> -<span class="sourceLineNo">482</span> count++;<a name="line.482"></a> -<span class="sourceLineNo">483</span><a name="line.483"></a> -<span class="sourceLineNo">484</span> // Using ByteBuffer for byte[] equality semantics<a name="line.484"></a> -<span class="sourceLineNo">485</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups = groupOrSplitPhase(table,<a name="line.485"></a> -<span class="sourceLineNo">486</span> pool, queue, startEndKeys);<a name="line.486"></a> -<span class="sourceLineNo">487</span><a name="line.487"></a> -<span class="sourceLineNo">488</span> if (!checkHFilesCountPerRegionPerFamily(regionGroups)) {<a name="line.488"></a> -<span class="sourceLineNo">489</span> // Error is logged inside checkHFilesCountPerRegionPerFamily.<a name="line.489"></a> -<span class="sourceLineNo">490</span> throw new IOException("Trying to load more than " + maxFilesPerRegionPerFamily<a name="line.490"></a> -<span class="sourceLineNo">491</span> + " hfiles to one family of one region");<a name="line.491"></a> -<span class="sourceLineNo">492</span> }<a name="line.492"></a> -<span class="sourceLineNo">493</span><a name="line.493"></a> -<span class="sourceLineNo">494</span> bulkLoadPhase(table, admin.getConnection(), pool, queue, regionGroups, copyFile);<a name="line.494"></a> +<span class="sourceLineNo">365</span> * @return List of filenames which were not found<a name="line.365"></a> +<span class="sourceLineNo">366</span> * @throws TableNotFoundException if table does not yet exist<a name="line.366"></a> +<span class="sourceLineNo">367</span> */<a name="line.367"></a> +<span class="sourceLineNo">368</span> public List<String> doBulkLoad(Map<byte[], List<Path>> map, final Admin admin, Table table,<a name="line.368"></a> +<span class="sourceLineNo">369</span> RegionLocator regionLocator, boolean silence, boolean copyFile)<a name="line.369"></a> +<span class="sourceLineNo">370</span> throws TableNotFoundException, IOException {<a name="line.370"></a> +<span class="sourceLineNo">371</span> if (!admin.isTableAvailable(regionLocator.getName())) {<a name="line.371"></a> +<span class="sourceLineNo">372</span> throw new TableNotFoundException("Table " + table.getName() + " is not currently available.");<a name="line.372"></a> +<span class="sourceLineNo">373</span> }<a name="line.373"></a> +<span class="sourceLineNo">374</span> // LQI queue does not need to be threadsafe -- all operations on this queue<a name="line.374"></a> +<span class="sourceLineNo">375</span> // happen in this thread<a name="line.375"></a> +<span class="sourceLineNo">376</span> Deque<LoadQueueItem> queue = new LinkedList<>();<a name="line.376"></a> +<span class="sourceLineNo">377</span> ExecutorService pool = null;<a name="line.377"></a> +<span class="sourceLineNo">378</span> SecureBulkLoadClient secureClient = null;<a name="line.378"></a> +<span class="sourceLineNo">379</span> try {<a name="line.379"></a> +<span class="sourceLineNo">380</span> prepareHFileQueue(map, table, queue, silence);<a name="line.380"></a> +<span class="sourceLineNo">381</span> if (queue.isEmpty()) {<a name="line.381"></a> +<span class="sourceLineNo">382</span> LOG.warn("Bulk load operation did not get any files to load");<a name="line.382"></a> +<span class="sourceLineNo">383</span> return null;<a name="line.383"></a> +<span class="sourceLineNo">384</span> }<a name="line.384"></a> +<span class="sourceLineNo">385</span> pool = createExecutorService();<a name="line.385"></a> +<span class="sourceLineNo">386</span> secureClient = new SecureBulkLoadClient(table.getConfiguration(), table);<a name="line.386"></a> +<span class="sourceLineNo">387</span> for (Map.Entry<byte[], List<Path>> entry : map.entrySet()) {<a name="line.387"></a> +<span class="sourceLineNo">388</span> for (Path p : entry.getValue()) {<a name="line.388"></a> +<span class="sourceLineNo">389</span> fs = p.getFileSystem(table.getConfiguration());<a name="line.389"></a> +<span class="sourceLineNo">390</span> break;<a name="line.390"></a> +<span class="sourceLineNo">391</span> }<a name="line.391"></a> +<span class="sourceLineNo">392</span> }<a name="line.392"></a> +<span class="sourceLineNo">393</span> return performBulkLoad(admin, table, regionLocator, queue, pool, secureClient, copyFile);<a name="line.393"></a> +<span class="sourceLineNo">394</span> } finally {<a name="line.394"></a> +<span class="sourceLineNo">395</span> cleanup(admin, queue, pool, secureClient);<a name="line.395"></a> +<span class="sourceLineNo">396</span> }<a name="line.396"></a> +<span class="sourceLineNo">397</span> }<a name="line.397"></a> +<span class="sourceLineNo">398</span><a name="line.398"></a> +<span class="sourceLineNo">399</span> /**<a name="line.399"></a> +<span class="sourceLineNo">400</span> * Perform a bulk load of the given directory into the given<a name="line.400"></a> +<span class="sourceLineNo">401</span> * pre-existing table. This method is not threadsafe.<a name="line.401"></a> +<span class="sourceLineNo">402</span> *<a name="line.402"></a> +<span class="sourceLineNo">403</span> * @param hfofDir the directory that was provided as the output path<a name="line.403"></a> +<span class="sourceLineNo">404</span> * of a job using HFileOutputFormat<a name="line.404"></a> +<span class="sourceLineNo">405</span> * @param admin the Admin<a name="line.405"></a> +<span class="sourceLineNo">406</span> * @param table the table to load into<a name="line.406"></a> +<span class="sourceLineNo">407</span> * @param regionLocator region locator<a name="line.407"></a> +<span class="sourceLineNo">408</span> * @param silence true to ignore unmatched column families<a name="line.408"></a> +<span class="sourceLineNo">409</span> * @param copyFile always copy hfiles if true<a name="line.409"></a> +<span class="sourceLineNo">410</span> * @throws TableNotFoundException if table does not yet exist<a name="line.410"></a> +<span class="sourceLineNo">411</span> */<a name="line.411"></a> +<span class="sourceLineNo">412</span> public void doBulkLoad(Path hfofDir, final Admin admin, Table table,<a name="line.412"></a> +<span class="sourceLineNo">413</span> RegionLocator regionLocator, boolean silence, boolean copyFile)<a name="line.413"></a> +<span class="sourceLineNo">414</span> throws TableNotFoundException, IOException {<a name="line.414"></a> +<span class="sourceLineNo">415</span> if (!admin.isTableAvailable(regionLocator.getName())) {<a name="line.415"></a> +<span class="sourceLineNo">416</span> throw new TableNotFoundException("Table " + table.getName() + " is not currently available.");<a name="line.416"></a> +<span class="sourceLineNo">417</span> }<a name="line.417"></a> +<span class="sourceLineNo">418</span><a name="line.418"></a> +<span class="sourceLineNo">419</span> /*<a name="line.419"></a> +<span class="sourceLineNo">420</span> * Checking hfile format is a time-consuming operation, we should have an option to skip<a name="line.420"></a> +<span class="sourceLineNo">421</span> * this step when bulkloading millions of HFiles. See HBASE-13985.<a name="line.421"></a> +<span class="sourceLineNo">422</span> */<a name="line.422"></a> +<span class="sourceLineNo">423</span> boolean validateHFile = getConf().getBoolean("hbase.loadincremental.validate.hfile", true);<a name="line.423"></a> +<span class="sourceLineNo">424</span> if (!validateHFile) {<a name="line.424"></a> +<span class="sourceLineNo">425</span> LOG.warn("You are skipping HFiles validation, it might cause some data loss if files " +<a name="line.425"></a> +<span class="sourceLineNo">426</span> "are not correct. If you fail to read data from your table after using this " +<a name="line.426"></a> +<span class="sourceLineNo">427</span> "option, consider removing the files and bulkload again without this option. " +<a name="line.427"></a> +<span class="sourceLineNo">428</span> "See HBASE-13985");<a name="line.428"></a> +<span class="sourceLineNo">429</span> }<a name="line.429"></a> +<span class="sourceLineNo">430</span> // LQI queue does not need to be threadsafe -- all operations on this queue<a name="line.430"></a> +<span class="sourceLineNo">431</span> // happen in this thread<a name="line.431"></a> +<span class="sourceLineNo">432</span> Deque<LoadQueueItem> queue = new LinkedList<>();<a name="line.432"></a> +<span class="sourceLineNo">433</span> ExecutorService pool = null;<a name="line.433"></a> +<span class="sourceLineNo">434</span> SecureBulkLoadClient secureClient = null;<a name="line.434"></a> +<span class="sourceLineNo">435</span> try {<a name="line.435"></a> +<span class="sourceLineNo">436</span> prepareHFileQueue(hfofDir, table, queue, validateHFile, silence);<a name="line.436"></a> +<span class="sourceLineNo">437</span><a name="line.437"></a> +<span class="sourceLineNo">438</span> if (queue.isEmpty()) {<a name="line.438"></a> +<span class="sourceLineNo">439</span> LOG.warn("Bulk load operation did not find any files to load in " +<a name="line.439"></a> +<span class="sourceLineNo">440</span> "directory " + hfofDir != null ? hfofDir.toUri() : "" + ". Does it contain files in " +<a name="line.440"></a> +<span class="sourceLineNo">441</span> "subdirectories that correspond to column family names?");<a name="line.441"></a> +<span class="sourceLineNo">442</span> return;<a name="line.442"></a> +<span class="sourceLineNo">443</span> }<a name="line.443"></a> +<span class="sourceLineNo">444</span> pool = createExecutorService();<a name="line.444"></a> +<span class="sourceLineNo">445</span> secureClient = new SecureBulkLoadClient(table.getConfiguration(), table);<a name="line.445"></a> +<span class="sourceLineNo">446</span> performBulkLoad(admin, table, regionLocator, queue, pool, secureClient, copyFile);<a name="line.446"></a> +<span class="sourceLineNo">447</span> } finally {<a name="line.447"></a> +<span class="sourceLineNo">448</span> cleanup(admin, queue, pool, secureClient);<a name="line.448"></a> +<span class="sourceLineNo">449</span> }<a name="line.449"></a> +<span class="sourceLineNo">450</span> }<a name="line.450"></a> +<span class="sourceLineNo">451</span><a name="line.451"></a> +<span class="sourceLineNo">452</span> List<String> performBulkLoad(final Admin admin, Table table, RegionLocator regionLocator,<a name="line.452"></a> +<span class="sourceLineNo">453</span> Deque<LoadQueueItem> queue, ExecutorService pool,<a name="line.453"></a> +<span class="sourceLineNo">454</span> SecureBulkLoadClient secureClient, boolean copyFile) throws IOException {<a name="line.454"></a> +<span class="sourceLineNo">455</span> int count = 0;<a name="line.455"></a> +<span class="sourceLineNo">456</span><a name="line.456"></a> +<span class="sourceLineNo">457</span> if(isSecureBulkLoadEndpointAvailable()) {<a name="line.457"></a> +<span class="sourceLineNo">458</span> LOG.warn("SecureBulkLoadEndpoint is deprecated. It will be removed in future releases.");<a name="line.458"></a> +<span class="sourceLineNo">459</span> LOG.warn("Secure bulk load has been integrated into HBase core.");<a name="line.459"></a> +<span class="sourceLineNo">460</span> }<a name="line.460"></a> +<span class="sourceLineNo">461</span><a name="line.461"></a> +<span class="sourceLineNo">462</span> //If using secure bulk load, get source delegation token, and<a name="line.462"></a> +<span class="sourceLineNo">463</span> //prepare staging directory and token<a name="line.463"></a> +<span class="sourceLineNo">464</span> // fs is the source filesystem<a name="line.464"></a> +<span class="sourceLineNo">465</span> fsDelegationToken.acquireDelegationToken(fs);<a name="line.465"></a> +<span class="sourceLineNo">466</span> bulkToken = secureClient.prepareBulkLoad(admin.getConnection());<a name="line.466"></a> +<span class="sourceLineNo">467</span> Pair<Multimap<ByteBuffer, LoadQueueItem>, List<String>> pair = null;<a name="line.467"></a> +<span class="sourceLineNo">468</span><a name="line.468"></a> +<span class="sourceLineNo">469</span> // Assumes that region splits can happen while this occurs.<a name="line.469"></a> +<span class="sourceLineNo">470</span> while (!queue.isEmpty()) {<a name="line.470"></a> +<span class="sourceLineNo">471</span> // need to reload split keys each iteration.<a name="line.471"></a> +<span class="sourceLineNo">472</span> final Pair<byte[][], byte[][]> startEndKeys = regionLocator.getStartEndKeys();<a name="line.472"></a> +<span class="sourceLineNo">473</span> if (count != 0) {<a name="line.473"></a> +<span class="sourceLineNo">474</span> LOG.info("Split occured while grouping HFiles, retry attempt " +<a name="line.474"></a> +<span class="sourceLineNo">475</span> + count + " with " + queue.size() + " files remaining to group or split");<a name="line.475"></a> +<span class="sourceLineNo">476</span> }<a name="line.476"></a> +<span class="sourceLineNo">477</span><a name="line.477"></a> +<span class="sourceLineNo">478</span> int maxRetries = getConf().getInt(HConstants.BULKLOAD_MAX_RETRIES_NUMBER, 10);<a name="line.478"></a> +<span class="sourceLineNo">479</span> maxRetries = Math.max(maxRetries, startEndKeys.getFirst().length + 1);<a name="line.479"></a> +<span class="sourceLineNo">480</span> if (maxRetries != 0 && count >= maxRetries) {<a name="line.480"></a> +<span class="sourceLineNo">481</span> throw new IOException("Retry attempted " + count +<a name="line.481"></a> +<span class="sourceLineNo">482</span> " times without completing, bailing out");<a name="line.482"></a> +<span class="sourceLineNo">483</span> }<a name="line.483"></a> +<span class="sourceLineNo">484</span> count++;<a name="line.484"></a> +<span class="sourceLineNo">485</span><a name="line.485"></a> +<span class="sourceLineNo">486</span> // Using ByteBuffer for byte[] equality semantics<a name="line.486"></a> +<span class="sourceLineNo">487</span> pair = groupOrSplitPhase(table, pool, queue, startEndKeys);<a name="line.487"></a> +<span class="sourceLineNo">488</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups = pair.getFirst();<a name="line.488"></a> +<span class="sourceLineNo">489</span><a name="line.489"></a> +<span class="sourceLineNo">490</span> if (!checkHFilesCountPerRegionPerFamily(regionGroups)) {<a name="line.490"></a> +<span class="sourceLineNo">491</span> // Error is logged inside checkHFilesCountPerRegionPerFamily.<a name="line.491"></a> +<span class="sourceLineNo">492</span> throw new IOException("Trying to load more than " + maxFilesPerRegionPerFamily<a name="line.492"></a> +<span class="sourceLineNo">493</span> + " hfiles to one family of one region");<a name="line.493"></a> +<span class="sourceLineNo">494</span> }<a name="line.494"></a> <span class="sourceLineNo">495</span><a name="line.495"></a> -<span class="sourceLineNo">496</span> // NOTE: The next iteration's split / group could happen in parallel to<a name="line.496"></a> -<span class="sourceLineNo">497</span> // atomic bulkloads assuming that there are splits and no merges, and<a name="line.497"></a> -<span class="sourceLineNo">498</span> // that we can atomically pull out the groups we want to retry.<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> if (!queue.isEmpty()) {<a name="line.501"></a> -<span class="sourceLineNo">502</span> throw new RuntimeException("Bulk load aborted with some files not yet loaded."<a name="line.502"></a> -<span class="sourceLineNo">503</span> + "Please check log for more details.");<a name="line.503"></a> -<span class="sourceLineNo">504</span> }<a name="line.504"></a> -<span class="sourceLineNo">505</span> }<a name="line.505"></a> -<span class="sourceLineNo">506</span><a name="line.506"></a> -<span class="sourceLineNo">507</span> /**<a name="line.507"></a> -<span class="sourceLineNo">508</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.508"></a> -<span class="sourceLineNo">509</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.509"></a> -<span class="sourceLineNo">510</span> * families in it.<a name="line.510"></a> -<span class="sourceLineNo">511</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.511"></a> -<span class="sourceLineNo">512</span> * @param table table to which hfiles should be loaded<a name="line.512"></a> -<span class="sourceLineNo">513</span> * @param queue queue which needs to be loaded into the table<a name="line.513"></a> -<span class="sourceLineNo">514</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.514"></a> -<span class="sourceLineNo">515</span> * @throws IOException If any I/O or network error occurred<a name="line.515"></a> -<span class="sourceLineNo">516</span> */<a name="line.516"></a> -<span class="sourceLineNo">517</span> public void prepareHFileQueue(Path hfilesDir, Table table, Deque<LoadQueueItem> queue,<a name="line.517"></a> -<span class="sourceLineNo">518</span> boolean validateHFile) throws IOException {<a name="line.518"></a> -<span class="sourceLineNo">519</span> prepareHFileQueue(hfilesDir, table, queue, validateHFile, false);<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> /**<a name="line.522"></a> -<span class="sourceLineNo">523</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.523"></a> -<span class="sourceLineNo">524</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.524"></a> -<span class="sourceLineNo">525</span> * families in it.<a name="line.525"></a> -<span class="sourceLineNo">526</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.526"></a> -<span class="sourceLineNo">527</span> * @param table table to which hfiles should be loaded<a name="line.527"></a> -<span class="sourceLineNo">528</span> * @param queue queue which needs to be loaded into the table<a name="line.528"></a> -<span class="sourceLineNo">529</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.529"></a> -<span class="sourceLineNo">530</span> * @param silence true to ignore unmatched column families<a name="line.530"></a> -<span class="sourceLineNo">531</span> * @throws IOException If any I/O or network error occurred<a name="line.531"></a> -<span class="sourceLineNo">532</span> */<a name="line.532"></a> -<span class="sourceLineNo">533</span> public void prepareHFileQueue(Path hfilesDir, Table table,<a name="line.533"></a> -<span class="sourceLineNo">534</span> Deque<LoadQueueItem> queue, boolean validateHFile, boolean silence) throws IOException {<a name="line.534"></a> -<span class="sourceLineNo">535</span> discoverLoadQueue(queue, hfilesDir, validateHFile);<a name="line.535"></a> -<span class="sourceLineNo">536</span> validateFamiliesInHFiles(table, queue, silence);<a name="line.536"></a> -<span class="sourceLineNo">537</span> }<a name="line.537"></a> -<span class="sourceLineNo">538</span><a name="line.538"></a> -<span class="sourceLineNo">539</span> /**<a name="line.539"></a> -<span class="sourceLineNo">540</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.540"></a> -<span class="sourceLineNo">541</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.541"></a> -<span class="sourceLineNo">542</span> * families in it.<a name="line.542"></a> -<span class="sourceLineNo">543</span> * @param map map of family to List of hfiles<a name="line.543"></a> -<span class="sourceLineNo">544</span> * @param table table to which hfiles should be loaded<a name="line.544"></a> -<span class="sourceLineNo">545</span> * @param queue queue which needs to be loaded into the table<a name="line.545"></a> -<span class="sourceLineNo">546</span> * @param silence true to ignore unmatched column families<a name="line.546"></a> -<span class="sourceLineNo">547</span> * @throws IOException If any I/O or network error occurred<a name="line.547"></a> -<span class="sourceLineNo">548</span> */<a name="line.548"></a> -<span class="sourceLineNo">549</span> public void prepareHFileQueue(Map<byte[], List<Path>> map, Table table,<a name="line.549"></a> -<span class="sourceLineNo">550</span> Deque<LoadQueueItem> queue, boolean silence) throws IOException {<a name="line.550"></a> -<span class="sourceLineNo">551</span> populateLoadQueue(queue, map);<a name="line.551"></a> -<span class="sourceLineNo">552</span> validateFamiliesInHFiles(table, queue, silence);<a name="line.552"></a> -<span class="sourceLineNo">553</span> }<a name="line.553"></a> -<span class="sourceLineNo">554</span><a name="line.554"></a> -<span class="sourceLineNo">555</span> // Initialize a thread pool<a name="line.555"></a> -<span class="sourceLineNo">556</span> private ExecutorService createExecutorService() {<a name="line.556"></a> -<span class="sourceLineNo">557</span> ThreadFactoryBuilder builder = new ThreadFactoryBuilder();<a name="line.557"></a> -<span class="sourceLineNo">558</span> builder.setNameFormat("LoadIncrementalHFiles-%1$d");<a name="line.558"></a> -<span class="sourceLineNo">559</span> ExecutorService pool = new ThreadPoolExecutor(nrThreads, nrThreads, 60, TimeUnit.SECONDS,<a name="line.559"></a> -<span class="sourceLineNo">560</span> new LinkedBlockingQueue<Runnable>(), builder.build());<a name="line.560"></a> -<span class="sourceLineNo">561</span> ((ThreadPoolExecutor) pool).allowCoreThreadTimeOut(true);<a name="line.561"></a> -<span class="sourceLineNo">562</span> return pool;<a name="line.562"></a> -<span class="sourceLineNo">563</span> }<a name="line.563"></a> -<span class="sourceLineNo">564</span><a name="line.564"></a> -<span class="sourceLineNo">565</span> /**<a name="line.565"></a> -<span class="sourceLineNo">566</span> * Checks whether there is any invalid family name in HFiles to be bulk loaded.<a name="line.566"></a> -<span class="sourceLineNo">567</span> */<a name="line.567"></a> -<span class="sourceLineNo">568</span> private void validateFamiliesInHFiles(Table table, Deque<LoadQueueItem> queue, boolean silence)<a name="line.568"></a> -<span class="sourceLineNo">569</span> throws IOException {<a name="line.569"></a> -<span class="sourceLineNo">570</span> Collection<HColumnDescriptor> families = table.getTableDescriptor().getFamilies();<a name="line.570"></a> -<span class="sourceLineNo">571</span> List<String> familyNames = new ArrayList<>(families.size());<a name="line.571"></a> -<span class="sourceLineNo">572</span> for (HColumnDescriptor family : families) {<a name="line.572"></a> -<span class="sourceLineNo">573</span> familyNames.add(family.getNameAsString());<a name="line.573"></a> -<span class="sourceLineNo">574</span> }<a name="line.574"></a> -<span class="sourceLineNo">575</span> Iterator<LoadQueueItem> queueIter = queue.iterator();<a name="line.575"></a> -<span class="sourceLineNo">576</span> while (queueIter.hasNext()) {<a name="line.576"></a> -<span class="sourceLineNo">577</span> LoadQueueItem lqi = queueIter.next();<a name="line.577"></a> -<span class="sourceLineNo">578</span> String familyNameInHFile = Bytes.toString(lqi.family);<a name="line.578"></a> -<span class="sourceLineNo">579</span> if (!familyNames.contains(familyNameInHFile)) {<a name="line.579"></a> -<span class="sourceLineNo">580</span> unmatchedFamilies.add(familyNameInHFile);<a name="line.580"></a> -<span class="sourceLineNo">581</span> }<a name="line.581"></a> -<span class="sourceLineNo">582</span> }<a name="line.582"></a> -<span class="sourceLineNo">583</span> if (unmatchedFamilies.size() > 0) {<a name="line.583"></a> -<span class="sourceLineNo">584</span> String msg =<a name="line.584"></a> -<span class="sourceLineNo">585</span> "Unmatched family names found: unmatched family names in HFiles to be bulkloaded: "<a name="line.585"></a> -<span class="sourceLineNo">586</span> + unmatchedFamilies + "; valid family names of table " + table.getName() + " are: "<a name="line.586"></a> -<span class="sourceLineNo">587</span> + familyNames;<a name="line.587"></a> -<span class="sourceLineNo">588</span> LOG.error(msg);<a name="line.588"></a> -<span class="sourceLineNo">589</span> if (!silence) throw new IOException(msg);<a name="line.589"></a> -<span class="sourceLineNo">590</span> }<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> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.594"></a> -<span class="sourceLineNo">595</span> * <ol><a name="line.595"></a> -<span class="sourceLineNo">596</span> * <li>LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)}</li><a name="line.596"></a> -<span class="sourceLineNo">597</span> * <li>LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)<a name="line.597"></a> -<span class="sourceLineNo">598</span> * </li><a name="line.598"></a> -<span class="sourceLineNo">599</span> * </ol><a name="line.599"></a> -<span class="sourceLineNo">600</span> * @param table Table to which these hfiles should be loaded to<a name="line.600"></a> -<span class="sourceLineNo">601</span> * @param conn Connection to use<a name="line.601"></a> -<span class="sourceLineNo">602</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.602"></a> -<span class="sourceLineNo">603</span> * @param startEndKeys starting and ending row keys of the region<a name="line.603"></a> -<span class="sourceLineNo">604</span> */<a name="line.604"></a> -<span class="sourceLineNo">605</span> public void loadHFileQueue(final Table table, final Connection conn, Deque<LoadQueueItem> queue,<a name="line.605"></a> -<span class="sourceLineNo">606</span> Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.606"></a> -<span class="sourceLineNo">607</span> loadHFileQueue(table, conn, queue, startEndKeys, false);<a name="line.607"></a> -<span class="sourceLineNo">608</span> }<a name="line.608"></a> -<span class="sourceLineNo">609</span><a name="line.609"></a> -<span class="sourceLineNo">610</span> /**<a name="line.610"></a> -<span class="sourceLineNo">611</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.611"></a> -<span class="sourceLineNo">612</span> * <ol><a name="line.612"></a> -<span class="sourceLineNo">613</span> * <li>LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)}</li><a name="line.613"></a> -<span class="sourceLineNo">614</span> * <li>LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)<a name="line.614"></a> -<span class="sourceLineNo">615</span> * </li><a name="line.615"></a> -<span class="sourceLineNo">616</span> * </ol><a name="line.616"></a> -<span class="sourceLineNo">617</span> * @param table Table to which these hfiles should be loaded to<a name="line.617"></a> -<span class="sourceLineNo">618</span> * @param conn Connection to use<a name="line.618"></a> -<span class="sourceLineNo">619</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.619"></a> -<span class="sourceLineNo">620</span> * @param startEndKeys starting and ending row keys of the region<a name="line.620"></a> -<span class="sourceLineNo">621</span> */<a name="line.621"></a> -<span class="sourceLineNo">622</span> public void loadHFileQueue(final Table table, final Connection conn, Deque<LoadQueueItem> queue,<a name="line.622"></a> -<span class="sourceLineNo">623</span> Pair<byte[][], byte[][]> startEndKeys, boolean copyFile) throws IOException {<a name="line.623"></a> -<span class="sourceLineNo">624</span> ExecutorService pool = null;<a name="line.624"></a> -<span class="sourceLineNo">625</span> try {<a name="line.625"></a> -<span class="sourceLineNo">626</span> pool = createExecutorService();<a name="line.626"></a> -<span class="sourceLineNo">627</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups =<a name="line.627"></a> -<span class="sourceLineNo">628</span> groupOrSplitPhase(table, pool, queue, startEndKeys);<a name="line.628"></a> -<span class="sourceLineNo">629</span> bulkLoadPhase(table, conn, pool, queue, regionGroups, copyFile);<a name="line.629"></a> -<span class="sourceLineNo">630</span> } finally {<a name="line.630"></a> -<span class="sourceLineNo">631</span> if (pool != null) {<a name="line.631"></a> -<span class="sourceLineNo">632</span> pool.shutdown();<a name="line.632"></a> -<span class="sourceLineNo">633</span> }<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> /**<a name="line.637"></a> -<span class="sourceLineNo">638</span> * This takes the LQI's grouped by likely regions and attempts to bulk load<a name="line.638"></a> -<span class="sourceLineNo">639</span> * them. Any failures are re-queued for another pass with the<a name="line.639"></a> -<span class="sourceLineNo">640</span> * groupOrSplitPhase.<a name="line.640"></a> -<span class="sourceLineNo">641</span> */<a name="line.641"></a> -<span class="sourceLineNo">642</span> protected void bulkLoadPhase(final Table table, final Connection conn,<a name="line.642"></a> -<span class="sourceLineNo">643</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.643"></a> -<span class="sourceLineNo">644</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups, boolean copyFile) throws IOException {<a name="line.644"></a> -<span class="sourceLineNo">645</span> // atomically bulk load the groups.<a name="line.645"></a> -<span class="sourceLineNo">646</span> Set<Future<List<LoadQueueItem>>> loadingFutures = new HashSet<>();<a name="line.646"></a> -<span class="sourceLineNo">647</span> for (Entry<ByteBuffer, ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()){<a name="line.647"></a> -<span class="sourceLineNo">648</span> final byte[] first = e.getKey().array();<a name="line.648"></a> -<span class="sourceLineNo">649</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.649"></a> -<span class="sourceLineNo">650</span><a name="line.650"></a> -<span class="sourceLineNo">651</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.651"></a> -<span class="sourceLineNo">652</span> @Override<a name="line.652"></a> -<span class="sourceLineNo">653</span> public List<LoadQueueItem> call() throws Exception {<a name="line.653"></a> -<span class="sourceLineNo">654</span> List<LoadQueueItem> toRetry =<a name="line.654"></a> -<span class="sourceLineNo">655</span> tryAtomicRegionLoad(conn, table.getName(), first, lqis, copyFile);<a name="line.655"></a> -<span class="sourceLineNo">656</span> return toRetry;<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> loadingFutures.add(pool.submit(call));<a name="line.659"></a> -<span class="sourceLineNo">660</span> }<a name="line.660"></a> -<span class="sourceLineNo">661</span><a name="line.661"></a> -<span class="sourceLineNo">662</span> // get all the results.<a name="line.662"></a> -<span class="sourceLineNo">663</span> for (Future<List<LoadQueueItem>> future : loadingFutures) {<a name="line.663"></a> -<span class="sourceLineNo">664</span> try {<a name="line.664"></a> -<span class="sourceLineNo">665</span> List<LoadQueueItem> toRetry = future.get();<a name="line.665"></a> -<span class="sourceLineNo">666</span><a name="line.666"></a> -<span class="sourceLineNo">667</span> // LQIs that are requeued to be regrouped.<a name="line.667"></a> -<span class="sourceLineNo">668</span> queue.addAll(toRetry);<a name="line.668"></a> -<span class="sourceLineNo">669</span><a name="line.669"></a> -<span class="sourceLineNo">670</span> } catch (ExecutionException e1) {<a name="line.670"></a> -<span class="sourceLineNo">671</span> Throwable t = e1.getCause();<a name="line.671"></a> -<span class="sourceLineNo">672</span> if (t instanceof IOException) {<a name="line.672"></a> -<span class="sourceLineNo">673</span> // At this point something unrecoverable has happened.<a name="line.673"></a> -<span class="sourceLineNo">674</span> // TODO Implement bulk load recovery<a name="line.674"></a> -<span class="sourceLineNo">675</span> throw new IOException("BulkLoad encountered an unrecoverable problem", t);<a name="line.675"></a> -<span class="sourceLineNo">676</span> }<a name="line.676"></a> -<span class="sourceLineNo">677</span> LOG.error("Unexpected execution exception during bulk load", e1);<a name="line.677"></a> -<span class="sourceLineNo">678</span> throw new IllegalStateException(t);<a name="line.678"></a> -<span class="sourceLineNo">679</span> } catch (InterruptedException e1) {<a name="line.679"></a> -<span class="sourceLineNo">680</span> LOG.error("Unexpected interrupted exception during bulk load", e1);<a name="line.680"></a> -<span class="sourceLineNo">681</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<a name="line.681"></a> -<span class="sourceLineNo">682</span> }<a name="line.682"></a> -<span class="sourceLineNo">683</span> }<a name="line.683"></a> -<span class="sourceLineNo">684</span> }<a name="line.684"></a> -<span class="sourceLineNo">685</span><a name="line.685"></a> -<span class="sourceLineNo">686</span> private boolean checkHFilesCountPerRegionPerFamily(<a name="line.686"></a> -<span class="sourceLineNo">687</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) {<a name="line.687"></a> -<span class="sourceLineNo">688</span> for (Entry<ByteBuffer,<a name="line.688"></a> -<span class="sourceLineNo">689</span> ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()) {<a name="line.689"></a> -<span class="sourceLineNo">690</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.690"></a> -<span class="sourceLineNo">691</span> HashMap<byte[], MutableInt> filesMap = new HashMap<>();<a name="line.691"></a> -<span class="sourceLineNo">692</span> for (LoadQueueItem lqi: lqis) {<a name="line.692"></a> -<span class="sourceLineNo">693</span> MutableInt count = filesMap.get(lqi.family);<a name="line.693"></a> -<span class="sourceLineNo">694</span> if (count == null) {<a name="line.694"></a> -<span class="sourceLineNo">695</span> count = new MutableInt();<a name="line.695"></a> -<span class="sourceLineNo">696</span> filesMap.put(lqi.family, count);<a name="line.696"></a> -<span class="sourceLineNo">697</span> }<a name="line.697"></a> -<span class="sourceLineNo">698</span> count.increment();<a name="line.698"></a> -<span class="sourceLineNo">699</span> if (count.intValue() > maxFilesPerRegionPerFamily) {<a name="line.699"></a> -<span class="sourceLineNo">700</span> LOG.error("Trying to load more than " + maxFilesPerRegionPerFamily<a name="line.700"></a> -<span class="sourceLineNo">701</span> + " hfiles to family " + Bytes.toStringBinary(lqi.family)<a name="line.701"></a> -<span class="sourceLineNo">702</span> + " of region with start key "<a name="line.702"></a> -<span class="sourceLineNo">703</span> + Bytes.toStringBinary(e.getKey()));<a name="line.703"></a> -<span class="sourceLineNo">704</span> return false;<a name="line.704"></a> -<span class="sourceLineNo">705</span> }<a name="line.705"></a> -<span class="sourceLineNo">706</span> }<a name="line.706"></a> -<span class="sourceLineNo">707</span> }<a name="line.707"></a> -<span class="sourceLineNo">708</span> return true;<a name="line.708"></a> -<span class="sourceLineNo">709</span> }<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> * @return A map that groups LQI by likely bulk load region targets.<a name="line.712"></a> -<span class="sourceLineNo">713</span> */<a name="line.713"></a> -<span class="sourceLineNo">714</span> private Multimap<ByteBuffer, LoadQueueItem> groupOrSplitPhase(final Table table,<a name="line.714"></a> -<span class="sourceLineNo">715</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.715"></a> -<span class="sourceLineNo">716</span> final Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.716"></a> -<span class="sourceLineNo">717</span> // <region start key, LQI> need synchronized only within this scope of this<a name="line.717"></a> -<span class="sourceLineNo">718</span> // phase because of the puts that happen in futures.<a name="line.718"></a> -<span class="sourceLineNo">719</span> Multimap<ByteBuffer, LoadQueueItem> rgs = HashMultimap.create();<a name="line.719"></a> -<span class="sourceLineNo">720</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups = Multimaps.synchronizedMultimap(rgs);<a name="line.720"></a> -<span class="sourceLineNo">721</span><a name="line.721"></a> -<span class="sourceLineNo">722</span> // drain LQIs and figure out bulk load groups<a name="line.722"></a> -<span class="sourceLineNo">723</span> Set<Future<List<LoadQueueItem>>> splittingFutures = new HashSet<>();<a name="line.723"></a> -<span class="sourceLineNo">724</span> while (!queue.isEmpty()) {<a name="line.724"></a> -<span class="sourceLineNo">725</span> final LoadQueueItem item = queue.remove();<a name="line.725"></a> -<span class="sourceLineNo">726</span><a name="line.726"></a> -<span class="sourceLineNo">727</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.727"></a> -<span class="sourceLineNo">728</span> @Override<a name="line.728"></a> -<span class="sourceLineNo">729</span> public List<LoadQueueItem> call() throws Exception {<a name="line.729"></a> -<span class="sourceLineNo">730</span> List<LoadQueueItem> splits = groupOrSplit(regionGroups, item, table, startEndKeys);<a name="line.730"></a> -<span class="sourceLineNo">731</span> return splits;<a name="line.731"></a> -<span class="sourceLineNo">732</span> }<a name="line.732"></a> -<span class="sourceLineNo">733</span> };<a name="line.733"></a> -<span class="sourceLineNo">734</span> splittingFutures.add(pool.submit(call));<a name="line.734"></a> -<span class="sourceLineNo">735</span> }<a name="line.735"></a> -<span class="sourceLineNo">736</span> // get all the results. All grouping and splitting must finish before<a name="line.736"></a> -<span class="sourceLineNo">737</span> // we can attempt the atomic loads.<a name="line.737"></a> -<span class="sourceLineNo">738</span> for (Future<List<LoadQueueItem>> lqis : splittingFutures) {<a name="line.738"></a> -<span class="sourceLineNo">739</span> try {<a name="line.739"></a> -<span class="sourceLineNo">740</span> List<LoadQueueItem> splits = lqis.get();<a name="line.740"></a> -<span class="sourceLineNo">741</span> if (splits != null) {<a name="line.741"></a> -<span class="sourceLineNo">742</span> queue.addAll(splits);<a name="line.742"></a> -<span class="sourceLineNo">743</span> }<a name="line.743"></a> -<span class="sourceLineNo">744</span> } catch (ExecutionException e1) {<a name="line.744"></a> -<span class="sourceLineNo">745</span> Throwable t = e1.getCause();<a name="line.745"></a> -<span class="sourceLineNo">746</span> if (t instanceof IOException) {<a name="line.746"></a> -<span class="sourceLineNo">747</span> LOG.error("IOException during splitting", e1);<a name="line.747"></a> -<span class="sourceLineNo">748</span> throw (IOException)t; // would have been thrown if not parallelized,<a name="line.748"></a> -<span class="sourceLineNo">749</span> }<a name="line.749"></a> -<span class="sourceLineNo">750</span> LOG.error("Unexpected execution exception during splitting", e1);<a name="line.750"></a> -<span class="sourceLineNo">751</span> throw new IllegalStateException(t);<a name="line.751"></a> -<span class="sourceLineNo">752</span> } catch (InterruptedException e1) {<a name="line.752"></a> -<span class="sourceLineNo">753</span> LOG.error("Unexpected interrupted exception during splitting", e1);<a name="line.753"></a> -<span class="sourceLineNo">754</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<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> return regionGroups;<a name="line.757"></a> -<span class="sourceLineNo">758</span> }<a name="line.758"></a> -<span class="sourceLineNo">759</span><a name="line.759"></a> -<span class="sourceLineNo">760</span> // unique file name for the table<a name="line.760"></a> -<span class="sourceLineNo">761</span> private String getUniqueName() {<a name="line.761"></a> -<span class="sourceLineNo">762</span> return UUID.randomUUID().toString().replaceAll("-", "");<a name="line.762"></a> -<span class="sourceLineNo">763</span> }<a name="line.763"></a> -<span class="sourceLineNo">764</span><a name="line.764"></a> -<span class="sourceLineNo">765</span> protected List<LoadQueueItem> splitStoreFile(final LoadQueueItem item,<a name="line.765"></a> -<span class="sourceLineNo">766</span> final Table table, byte[] startKey,<a name="line.766"></a> -<span class="sourceLineNo">767</span> byte[] splitKey) throws IOException {<a name="line.767"></a> -<span class="sourceLineNo">768</span> final Path hfilePath = item.hfilePath;<a name="line.768"></a> -<span class="sourceLineNo">769</span><a name="line.769"></a> -<span class="sourceLineNo">770</span> Path tmpDir = item.hfilePath.getParent();<a name="line.770"></a> -<span class="sourceLineNo">771</span> if (!tmpDir.getName().equals(TMP_DIR)) {<a name="line.771"></a> -<span class="sourceLineNo">772</span> tmpDir = new Path(tmpDir, TMP_DIR);<a name="line.772"></a> +<span class="sourceLineNo">496</span> bulkLoadPhase(table, admin.getConnection(), pool, queue, regionGroups, copyFile);<a name="line.496"></a> +<span class="sourceLineNo">497</span><a name="line.497"></a> +<span class="sourceLineNo">498</span> // NOTE: The next iteration's split / group could happen in parallel to<a name="line.498"></a> +<span class="sourceLineNo">499</span> // atomic bulkloads assuming that there are splits and no merges, and<a name="line.499"></a> +<span class="sourceLineNo">500</span> // that we can atomically pull out the groups we want to retry.<a name="line.500"></a> +<span class="sourceLineNo">501</span> }<a name="line.501"></a> +<span class="sourceLineNo">502</span><a name="line.502"></a> +<span class="sourceLineNo">503</span> if (!queue.isEmpty()) {<a name="line.503"></a> +<span class="sourceLineNo">504</span> throw new RuntimeException("Bulk load aborted with some files not yet loaded."<a name="line.504"></a> +<span class="sourceLineNo">505</span> + "Please check log for more details.");<a name="line.505"></a> +<span class="sourceLineNo">506</span> }<a name="line.506"></a> +<span class="sourceLineNo">507</span> if (pair == null) return null;<a name="line.507"></a> +<span class="sourceLineNo">508</span> return pair.getSecond();<a name="line.508"></a> +<span class="sourceLineNo">509</span> }<a name="line.509"></a> +<span class="sourceLineNo">510</span><a name="line.510"></a> +<span class="sourceLineNo">511</span> /**<a name="line.511"></a> +<span class="sourceLineNo">512</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.512"></a> +<span class="sourceLineNo">513</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.513"></a> +<span class="sourceLineNo">514</span> * families in it.<a name="line.514"></a> +<span class="sourceLineNo">515</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.515"></a> +<span class="sourceLineNo">516</span> * @param table table to which hfiles should be loaded<a name="line.516"></a> +<span class="sourceLineNo">517</span> * @param queue queue which needs to be loaded into the table<a name="line.517"></a> +<span class="sourceLineNo">518</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.518"></a> +<span class="sourceLineNo">519</span> * @throws IOException If any I/O or network error occurred<a name="line.519"></a> +<span class="sourceLineNo">520</span> */<a name="line.520"></a> +<span class="sourceLineNo">521</span> public void prepareHFileQueue(Path hfilesDir, Table table, Deque<LoadQueueItem> queue,<a name="line.521"></a> +<span class="sourceLineNo">522</span> boolean validateHFile) throws IOException {<a name="line.522"></a> +<span class="sourceLineNo">523</span> prepareHFileQueue(hfilesDir, table, queue, validateHFile, false);<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> /**<a name="line.526"></a> +<span class="sourceLineNo">527</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.527"></a> +<span class="sourceLineNo">528</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.528"></a> +<span class="sourceLineNo">529</span> * families in it.<a name="line.529"></a> +<span class="sourceLineNo">530</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.530"></a> +<span class="sourceLineNo">531</span> * @param table table to which hfiles should be loaded<a name="line.531"></a> +<span class="sourceLineNo">532</span> * @param queue queue which needs to be loaded into the table<a name="line.532"></a> +<span class="sourceLineNo">533</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.533"></a> +<span class="sourceLineNo">534</span> * @param silence true to ignore unmatched column families<a name="line.534"></a> +<span class="sourceLineNo">535</span> * @throws IOException If any I/O or network error occurred<a name="line.535"></a> +<span class="sourceLineNo">536</span> */<a name="line.536"></a> +<span class="sourceLineNo">537</span> public void prepareHFileQueue(Path hfilesDir, Table table,<a name="line.537"></a> +<span class="sourceLineNo">538</span> Deque<LoadQueueItem> queue, boolean validateHFile, boolean silence) throws IOException {<a name="line.538"></a> +<span class="sourceLineNo">539</span> discoverLoadQueue(queue, hfilesDir, validateHFile);<a name="line.539"></a> +<span class="sourceLineNo">540</span> validateFamiliesInHFiles(table, queue, silence);<a name="line.540"></a> +<span class="sourceLineNo">541</span> }<a name="line.541"></a> +<span class="sourceLineNo">542</span><a name="line.542"></a> +<span class="sourceLineNo">543</span> /**<a name="line.543"></a> +<span class="sourceLineNo">544</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.544"></a> +<span class="sourceLineNo">545</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.545"></a> +<span class="sourceLineNo">546</span> * families in it.<a name="line.546"></a> +<span class="sourceLineNo">547</span> * @param map map of family to List of hfiles<a name="line.547"></a> +<span class="sourceLineNo">548</span> * @param table table to which hfiles should be loaded<a name="line.548"></a> +<span class="sourceLineNo">549</span> * @param queue queue which needs to be loaded into the table<a name="line.549"></a> +<span class="sourceLineNo">550</span> * @param silence true to ignore unmatched column families<a name="line.550"></a> +<span class="sourceLineNo">551</span> * @throws IOException If any I/O or network error occurred<a name="line.551"></a> +<span class="sourceLineNo">552</span> */<a name="line.552"></a> +<span class="sourceLineNo">553</span> public void prepareHFileQueue(Map<byte[], List<Path>> map, Table table,<a name="line.553"></a> +<span class="sourceLineNo">554</span> Deque<LoadQueueItem> queue, boolean silence) throws IOException {<a name="line.554"></a> +<span class="sourceLineNo">555</span> populateLoadQueue(queue, map);<a name="line.555"></a> +<span class="sourceLineNo">556</span> validateFamiliesInHFiles(table, queue, silence);<a name="line.556"></a> +<span class="sourceLineNo">557</span> }<a name="line.557"></a> +<span class="sourceLineNo">558</span><a name="line.558"></a> +<span class="sourceLineNo">559</span> // Initialize a thread pool<a name="line.559"></a> +<span class="sourceLineNo">560</span> private ExecutorService createExecutorService() {<a name="line.560"></a> +<span class="sourceLineNo">561</span> ThreadFactoryBuilder builder = new ThreadFactoryBuilder();<a name="line.561"></a> +<span class="sourceLineNo">562</span> builder.setNameFormat("LoadIncrementalHFiles-%1$d");<a name="line.562"></a> +<span class="sourceLineNo">563</span> ExecutorService pool = new ThreadPoolExecutor(nrThreads, nrThreads, 60, TimeUnit.SECONDS,<a name="line.563"></a> +<span class="sourceLineNo">564</span> new LinkedBlockingQueue<Runnable>(), builder.build());<a name="line.564"></a> +<span class="sourceLineNo">565</span> ((ThreadPoolExecutor) pool).allowCoreThreadTimeOut(true);<a name="line.565"></a> +<span class="sourceLineNo">566</span> return pool;<a name="line.566"></a> +<span class="sourceLineNo">567</span> }<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> * Checks whether there is any invalid family name in HFiles to be bulk loaded.<a name="line.570"></a> +<span class="sourceLineNo">571</span> */<a name="line.571"></a> +<span class="sourceLineNo">572</span> private void validateFamiliesInHFiles(Table table, Deque<LoadQueueItem> queue, boolean silence)<a name="line.572"></a> +<span class="sourceLineNo">573</span> throws IOException {<a name="line.573"></a> +<span class="sourceLineNo">574</span> Collection<HColumnDescriptor> families = table.getTableDescriptor().getFamilies();<a name="line.574"></a> +<span class="sourceLineNo">575</span> List<String> familyNames = new ArrayList<>(families.size());<a name="line.575"></a> +<span class="sourceLineNo">576</span> for (HColumnDescriptor family : families) {<a name="line.576"></a> +<span class="sourceLineNo">577</span> familyNames.add(family.getNameAsString());<a name="line.577"></a> +<span class="sourceLineNo">578</span> }<a name="line.578"></a> +<span class="sourceLineNo">579</span> Iterator<LoadQueueItem> queueIter = queue.iterator();<a name="line.579"></a> +<span class="sourceLineNo">580</span> while (queueIter.hasNext()) {<a name="line.580"></a> +<span class="sourceLineNo">581</span> LoadQueueItem lqi = queueIter.next();<a name="line.581"></a> +<span class="sourceLineNo">582</span> String familyNameInHFile = Bytes.toString(lqi.family);<a name="line.582"></a> +<span class="sourceLineNo">583</span> if (!familyNames.contains(familyNameInHFile)) {<a name="line.583"></a> +<span class="sourceLineNo">584</span> unmatchedFamilies.add(familyNameInHFile);<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 (unmatchedFamilies.size() > 0) {<a name="line.587"></a> +<span class="sourceLineNo">588</span> String msg =<a name="line.588"></a> +<span class="sourceLineNo">589</span> "Unmatched family names found: unmatched family names in HFiles to be bulkloaded: "<a name="line.589"></a> +<span class="sourceLineNo">590</span> + unmatchedFamilies + "; valid family names of table " + table.getName() + " are: "<a name="line.590"></a> +<span class="sourceLineNo">591</span> + familyNames;<a name="line.591"></a> +<span class="sourceLineNo">592</span> LOG.error(msg);<a name="line.592"></a> +<span class="sourceLineNo">593</span> if (!silence) throw new IOException(msg);<a name="line.593"></a> +<span class="sourceLineNo">594</span> }<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> /**<a name="line.597"></a> +<span class="sourceLineNo">598</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.598"></a> +<span class="sourceLineNo">599</span> * <ol><a name="line.599"></a> +<span class="sourceLineNo">600</span> * <li>LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)}</li><a name="line.600"></a> +<span class="sourceLineNo">601</span> * <li>LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)<a name="line.601"></a> +<span class="sourceLineNo">602</span> * </li><a name="line.602"></a> +<span class="sourceLineNo">603</span> * </ol><a name="line.603"></a> +<span class="sourceLineNo">604</span> * @param table Table to which these hfiles should be loaded to<a name="line.604"></a> +<span class="sourceLineNo">605</span> * @param conn Connection to use<a name="line.605"></a> +<span class="sourceLineNo">606</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.606"></a> +<span class="sourceLineNo">607</span> * @param startEndKeys starting and ending row keys of the region<a name="line.607"></a> +<span class="sourceLineNo">608</span> */<a name="line.608"></a> +<span class="sourceLineNo">609</span> public void loadHFileQueue(final Table table, final Connection conn, Deque<LoadQueueItem> queue,<a name="line.609"></a> +<span class="sourceLineNo">610</span> Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.610"></a> +<span class="sourceLineNo">611</span> loadHFileQueue(table, conn, queue, startEndKeys, false);<a name="line.611"></a> +<span class="sourceLineNo">612</span> }<a name="line.612"></a> +<span class="sourceLineNo">613</span><a name="line.613"></a> +<span class="sourceLineNo">614</span> /**<a name="line.614"></a> +<span class="sourceLineNo">615</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.615"></a> +<span class="sourceLineNo">616</span> * <ol><a name="line.616"></a> +<span class="sourceLineNo">617</span> * <li>LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)}</li><a name="line.617"></a> +<span class="sourceLineNo">618</span> * <li>LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)<a name="line.618"></a> +<span class="sourceLineNo">619</span> * </li><a name="line.619"></a> +<span class="sourceLineNo">620</span> * </ol><a name="line.620"></a> +<span class="sourceLineNo">621</span> * @param table Table to which these hfiles should be loaded to<a name="line.621"></a> +<span class="sourceLineNo">622</span> * @param conn Connection to use<a name="line.622"></a> +<span class="sourceLineNo">623</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.623"></a> +<span class="sourceLineNo">624</span> * @param startEndKeys starting and ending row keys of the region<a name="line.624"></a> +<span class="sourceLineNo">625</span> */<a name="line.625"></a> +<span class="sourceLineNo">626</span> public void loadHFileQueue(final Table table, final Connection conn, Deque<LoadQueueItem> queue,<a name="line.626"></a> +<span class="sourceLineNo">627</span> Pair<byte[][], byte[][]> startEndKeys, boolean copyFile) throws IOException {<a name="line.627"></a> +<span class="sourceLineNo">628</span> ExecutorService pool = null;<a name="line.628"></a> +<span class="sourceLineNo">629</span> try {<a name="line.629"></a> +<span class="sourceLineNo">630</span> pool = createExecutorService();<a name="line.630"></a> +<span class="sourceLineNo">631</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups =<a name="line.631"></a> +<span class="sourceLineNo">632</span> groupOrSplitPhase(table, pool, queue, startEndKeys).getFirst();<a name="line.632"></a> +<span class="sourceLineNo">633</span> bulkLoadPhase(table, conn, pool, queue, regionGroups, copyFile);<a name="line.633"></a> +<span class="sourceLineNo">634</span> } finally {<a name="line.634"></a> +<span class="sourceLineNo">635</span> if (pool != null) {<a name="line.635"></a> +<span class="sourceLineNo">636</span> pool.shutdown();<a name="line.636"></a> +<span class="sourceLineNo">637</span> }<a name="line.637"></a> +<span class="sourceLineNo">638</span> }<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> /**<a name="line.641"></a> +<span class="sourceLineNo">642</span> * This takes the LQI's grouped by likely regions and attempts to bulk load<a name="line.642"></a> +<span class="sourceLineNo">643</span> * them. Any failures are re-queued for another pass with the<a name="line.643"></a> +<span class="sourceLineNo">644</span> * groupOrSplitPhase.<a name="line.644"></a> +<span class="sourceLineNo">645</span> */<a name="line.645"></a> +<span class="sourceLineNo">646</span> protected void bulkLoadPhase(final Table table, final Connection conn,<a name="line.646"></a> +<span class="sourceLineNo">647</span> ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.647"></a> +<span class="sourceLineNo">648</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups, boolean copyFile) throws IOException {<a name="line.648"></a> +<span class="sourceLineNo">649</span> // atomically bulk load the groups.<a name="line.649"></a> +<span class="sourceLineNo">650</span> Set<Future<List<LoadQueueItem>>> loadingFutures = new HashSet<>();<a name="line.650"></a> +<span class="sourceLineNo">651</span> for (Entry<ByteBuffer, ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()){<a name="line.651"></a> +<span class="sourceLineNo">652</span> final byte[] first = e.getKey().array();<a name="line.652"></a> +<span class="sourceLineNo">653</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.653"></a> +<span class="sourceLineNo">654</span><a name="line.654"></a> +<span class="sourceLineNo">655</span> final Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.655"></a> +<span class="sourceLineNo">656</span> @Override<a name="line.656"></a> +<span class="sourceLineNo">657</span> public List<LoadQueueItem> call() throws Exception {<a name="line.657"></a> +<span class="sourceLineNo">658</span> List<LoadQueueItem> toRetry =<a name="line.658"></a> +<span class="sourceLineNo">659</span> tryAtomicRegionLoad(conn, table.getName(), first, lqis, copyFile);<a name="line.659"></a> +<span class="sourceLineNo">660</span> return toRetry;<a name="line.660"></a> +<span class="sourceLineNo">661</span> }<a name="line.661"></a> +<span class="sourceLineNo">662</span> };<a name="line.662"></a> +<span class="sourceLineNo">663</span> loadingFutures.add(pool.submit(call));<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> // get all the results.<a name="line.666"></a> +<span class="sourceLineNo">667</span> for (Future<List<LoadQueueItem>> future : loadingFutures) {<a name="line.667"></a> +<span class="sourceLineNo">668</span> try {<a name="line.668"></a> +<span class="sourceLineNo">669</span> List<LoadQueueItem> toRetry = future.get();<a name="line.669"></a> +<span class="sourceLineNo">670</span><a name="line.670"></a> +<span class="sourceLineNo">671</span> // LQIs that are requeued to be regrouped.<a name="line.671"></a> +<span class="sourceLineNo">672</span> queue.addAll(toRetry);<a name="line.672"></a> +<span class="sourceLineNo">673</span><a name="line.673"></a> +<span class="sourceLineNo">674</span> } catch (ExecutionException e1) {<a name="line.674"></a> +<span class="sourceLineNo">675</span> Throwable t = e1.getCause();<a name="line.675"></a> +<span class="sourceLineNo">676</span> if (t instanceof IOException) {<a name="line.676"></a> +<span class="sourceLineNo">677</span> // At this point something unrecoverable has happened.<a name="line.677"></a> +<span class="sourceLineNo">678</span> // TODO Implement bulk load recovery<a name="line.678"></a> +<span class="sourceLineNo">679</span> throw new IOException("BulkLoad encountered an unrecoverable problem", t);<a name="line.679"></a> +<span class="sourceLineNo">680</span> }<a name="line.680"></a> +<span class="sourceLineNo">681</span> LOG.error("Unexpected execution exception during bulk load", e1);<a name="line.681"></a> +<span class="sourceLineNo">682</span> throw new IllegalStateException(t);<a name="line.682"></a> +<span class="sourceLineNo">683</span> } catch (InterruptedException e1) {<a name="line.683"></a> +<span class="sourceLineNo">684</span> LOG.error("Unexpected interrupted exception during bulk load", e1);<a name="line.684"></a> +<span class="sourceLineNo">685</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<a name="line.685"></a> +<span class="sourceLineNo">686</span> }<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> private boolean checkHFilesCountPerRegionPerFamily(<a name="line.690"></a> +<span class="sourceLineNo">691</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) {<a name="line.691"></a> +<span class="sourceLineNo">692</span> for (Entry<ByteBuffer,<a name="line.692"></a> +<span class="sourceLineNo">693</span> ? extends Collection<LoadQueueItem>> e: regionGroups.asMap().entrySet()) {<a name="line.693"></a> +<span class="sourceLineNo">694</span> final Collection<LoadQueueItem> lqis = e.getValue();<a name="line.694"></a> +<span class="sourceLineNo">695</span> HashMap<byte[], MutableInt> filesMap = new HashMap<>();<a name="line.695"></a> +<span class="sourceLineNo">696</span> for (LoadQueueItem lqi: lqis) {<a name="line.696"></a> +<span class="sourceLineNo">697</span> MutableInt count = filesMap.get(lqi.family);<a name="line.697"></a> +<span class="sourceLineNo">698</span> if (count == null) {<a name="line.698"></a> +<span class="sourceLineNo">699</span> count = new MutableInt();<a name="line.699"></a> +<span class="sourceLineNo">700</span> filesMap.put(lqi.family, count);<a name="line.700"></a> +<span class="sourceLineNo">701</span> }<a name="line.701"></a> +<span class="sourceLineNo">702</span> count.increment();<a name="line.702"></a> +<span class="sourceLineNo">703</span> if (count.intValue() > maxFilesPerRegionPerFamily) {<a name="line.703"></a> +<span class="sourceLineNo">704</span> LOG.error("Trying to load more than " + maxFilesPerRegionPerFamily<a name="line.704"></a> +<span class="sourceLineNo">705</span> + " hfiles to family " + Bytes.toStringBinary(lqi.family)<a name="line.705"></a> +<span class="sourceLineNo">706</span> + " of region with start key "<a name="line.706"></a> +<span class="sourceLineNo">707</span> + Bytes.toStringBinary(e.getKey()));<a name="line.707"></a> +<span class="sourceLineNo">708</span> return false;<a name="line.708"></a> +<span class="sourceLineNo">709</span> }<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> return true;<a name="line.712"></a> +<span class="sourceLineNo">713</span> }<a name="line.713"></a> +<span class="sourceLineNo">714</span><a name="line.714"></a> +<span class="sourceLineNo">715</span> /**<a name="line.715"></a> +<span class="sourceLineNo">716</span> * @param table the table to load into<a name="line.716"></a> +<span class="sourceLineNo">717</span> * @param pool the ExecutorService<a name="line.717"></a> +<span class="sourceLineNo">718</span> * @param queue the queue for LoadQueueItem<a name="line.718"></a> +<span class="sourceLineNo">719</span> * @param startEndKeys start and end keys<a name="line.719"></a> +<span class="sourceLineNo">720</span> * @return A map that groups LQI by likely bulk load region targets and List of missing hfiles.<a name="line.720"></a> +<span class="sourceLineNo">721</span> */<a name="line.721"></a> +<span class="sourceLineNo">722</span> private Pair<Multimap<ByteBuffer, LoadQueueItem>, List<String>> groupOrSplitPhase(<a name="line.722"></a> +<span class="sourceLineNo">723</span> final Table table, ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.723"></a> +<span class="sourceLineNo">724</span> final Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.724"></a> +<span class="sourceLineNo">725</span> // <region start key, LQI> need synchronized only within this scope of this<a name="line.725"></a> +<span class="sourceLineNo">726</span> // phase because of the puts that happen in futures.<a name="line.726"></a> +<span class="sourceLineNo">727</span> Multimap<ByteBuffer, LoadQueueItem> rgs = HashMultimap.create();<a name="line.727"></a> +<span class="sourceLineNo">728</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups = Multimaps.synchronizedMultimap(rgs);<a name="line.728"></a> +<span class="sourceLineNo">729</span> List<String> missingHFiles = new ArrayList<>();<a name="line.729"></a> +<span class="sourceLineNo">730</span> Pair<Multimap<ByteBuffer, LoadQueueItem>, List<String>> pair = new Pair<>(regionGroups,<a name="line.730"></a> +<span class="sourceLineNo">731</span> missingHFiles);<a name="line.731"></a> +<span class="sourceLineNo">732</span><a name="line.732"></a> +<span class="sourceLineNo">733</span> // drain LQIs and figure out bulk load groups<a name="line.733"></a> +<span class="sourceLineNo">734</span> Set<Future<Pair<List<LoadQueueItem>, String>>> splittingFutures = new HashSet<>();<a name="line.734"></a> +<span class="sourceLineNo">735</span> while (!queue.isEmpty()) {<a name="line.735"></a> +<span class="sourceLineNo">736</span> final LoadQueueItem item = queue.remove();<a name="line.736"></a> +<span class="sourceLineNo">737</span><a name="line.737"></a> +<span class="sourceLineNo">738</span> final Callable<Pair<List<LoadQueueItem>, String>> call =<a name="line.738"></a> +<span class="sourceLineNo">739</span> new Callable<Pair<List<LoadQueueItem>, String>>() {<a name="line.739"></a> +<span class="sourceLineNo">740</span> @Override<a name="line.740"></a> +<span class="sourceLineNo">741</span> public Pair<List<LoadQueueItem>, String> call() throws Exception {<a name="line.741"></a> +<span class="sourceLineNo">742</span> Pair<List<LoadQueueItem>, String> splits = groupOrSplit(regionGroups, item, table,<a name="line.742"></a> +<span class="sourceLineNo">743</span> startEndKeys);<a name="line.743"></a> +<span class="sourceLineNo">744</span> return splits;<a name="line.744"></a> +<span class="sourceLineNo">745</span> }<a name="line.745"></a> +<span class="sourceLineNo">746</span> };<a name="line.746"></a> +<span class="sourceLineNo">747</span> splittingFutures.add(pool.submit(call));<a name="line.747"></a> +<span class="sourceLineNo">748</span> }<a name="line.748"></a> +<span class="sourceLineNo">749</span> // get all the results. All grouping and splitting must finish before<a name="line.749"></a> +<span class="sourceLineNo">750</span> // we can attempt the atomic loads.<a name="line.750"></a> +<span class="sourceLineNo">751</span> for (Future<Pair<List<LoadQueueItem>, String>> lqis : splittingFutures) {<a name="line.751"></a> +<span class="sourceLineNo">752</span> try {<a name="line.752"></a> +<span class="sourceLineNo">753</span> Pair<List<LoadQueueItem>, String> splits = lqis.get();<a name="line.753"></a> +<span class="sourceLineNo">754</span> if (splits != null) {<a name="line.754"></a> +<span class="sourceLineNo">755</span> if (splits.getFirst() != null) {<a name="line.755"></a> +<span class="sourceLineNo">756</span> queue.addAll(splits.getFirst());<a name="line.756"></a> +<span class="sourceLineNo">757</span> } else {<a name="line.757"></a> +<span class="sourceLineNo">758</span> missingHFiles.add(splits.getSecond());<a name="line.758"></a> +<span class="sourceLineNo">759</span> }<a name="line.759"></a> +<span class="sourceLineNo">760</span> }<a name="line.760"></a> +<span class="sourceLineNo">761</span> } catch (ExecutionException e1) {<a name="line.761"></a> +<span class="sourceLineNo">762</span> Throwable t = e1.getCause();<a name="line.762"></a> +<span class="sourceLineNo">763</span> if (t instanceof IOException) {<a name="line.763"></a> +<span class="sourceLineNo">764</span> LOG.error("IOException during splitting", e1);<a name="line.764"></a> +<span class="sourceLineNo">765</span> throw (IOException)t; // would have been thrown if not parallelized,<a name="line.765"></a> +<span class="sourceLineNo">766</span> }<a name="line.766"></a> +<span class="sourceLineNo">767</span> LOG.error("Unexpected execution exception during splitting", e1);<a name="line.767"></a> +<span class="sourceLineNo">768</span> throw new IllegalStateException(t);<a name="line.768"></a> +<span class="sourceLineNo">769</span> } catch (InterruptedException e1) {<a name="line.769"></a> +<span class="sourceLineNo">770</span> LOG.error("Unexpected interrupted exception during splitting", e1);<a name="line.770"></a> +<span class="sourceLineNo">771</span> throw (InterruptedIOException)new InterruptedIOException().initCause(e1);<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> LOG.info("HFile at " + hfilePath + " no longer fits inside a single " +<a name="line.775"></a> -<span class="sourceLineNo">776</span> "region. Splitting...");<a name="line.776"></a> -<span class="sourceLineNo">777</span><a name="line.777"></a> -<span class="sourceLineNo">778</span> String uniqueName = getUniqueName();<a name="line.778"></a> -<span class="sourceLineNo">779</span> HColumnDescriptor familyDesc = table.getTableDescriptor().getFamily(item.family);<a name="line.779"></a> -<span class="sourceLineNo">780</span><a name="line.780"></a> -<span class="sourceLineNo">781</span> Path botOut = new Path(tmpDir, uniqueName + ".bottom");<a name="line.781"></a> -<span class="sourceLineNo">782</span> Path topOut = new Path(tmpDir, uniqueName + ".top");<a name="line.782"></a> -<span class="sourceLineNo">783</span> splitStoreFile(getConf(), hfilePath, familyDesc, splitKey, botOut, topOut);<a name="line.783"></a> -<span class="sourceLineNo">784</span><a name="line.784"></a> -<span class="sourceLineNo">785</span> FileSystem fs = tmpDir.getFileSystem(getConf());<a name="line.785"></a> -<span class="sourceLineNo">786</span> fs.setPermission(tmpDir, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.786"></a> -<span class="sourceLineNo">787</span> fs.setPermission(botOut, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.787"></a> -<span class="sourceLineNo">788</span> fs.setPermission(topOut, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.788"></a> -<span class="sourceLineNo">789</span><a name="line.789"></a> -<span class="sourceLineNo">790</span> // Add these back at the *front* of the queue, so there's a lower<a name="line.790"></a> -<span class="s
<TRUNCATED>
