Examples with SegmentStatusManager org.apache.carbondata.core.statusmanager.SegmentStatusManager used on opensource projects

Example 1 with SegmentStatusManager

use of org.apache.carbondata.core.statusmanager.SegmentStatusManager in project carbondata by apache.

the class CarbonUpdateUtil method cleanUpDeltaFiles.

/**
   * Handling of the clean up of old carbondata files, index files , delte delta,
   * update status files.
   * @param table clean up will be handled on this table.
   * @param forceDelete if true then max query execution timeout will not be considered.
   */
public static void cleanUpDeltaFiles(CarbonTable table, boolean forceDelete) {
    SegmentStatusManager ssm = new SegmentStatusManager(table.getAbsoluteTableIdentifier());
    CarbonTablePath carbonTablePath = CarbonStorePath.getCarbonTablePath(table.getAbsoluteTableIdentifier().getStorePath(), table.getAbsoluteTableIdentifier().getCarbonTableIdentifier());
    LoadMetadataDetails[] details = ssm.readLoadMetadata(table.getMetaDataFilepath());
    String validUpdateStatusFile = "";
    for (LoadMetadataDetails segment : details) {
        // take the update status file name from 0th segment.
        validUpdateStatusFile = ssm.getUpdateStatusFileName(details);
        if (segment.getLoadStatus().equalsIgnoreCase(CarbonCommonConstants.STORE_LOADSTATUS_SUCCESS) || segment.getLoadStatus().equalsIgnoreCase(CarbonCommonConstants.STORE_LOADSTATUS_PARTIAL_SUCCESS)) {
            // take the list of files from this segment.
            String segmentPath = carbonTablePath.getCarbonDataDirectoryPath("0", segment.getLoadName());
            CarbonFile segDir = FileFactory.getCarbonFile(segmentPath, FileFactory.getFileType(segmentPath));
            CarbonFile[] allSegmentFiles = segDir.listFiles();
            // scan through the segment and find the carbondatafiles and index files.
            SegmentUpdateStatusManager updateStatusManager = new SegmentUpdateStatusManager(table.getAbsoluteTableIdentifier());
            // get Invalid update  delta files.
            CarbonFile[] invalidUpdateDeltaFiles = updateStatusManager.getUpdateDeltaFilesList(segment.getLoadName(), false, CarbonCommonConstants.UPDATE_DELTA_FILE_EXT, true, allSegmentFiles);
            for (CarbonFile invalidFile : invalidUpdateDeltaFiles) {
                compareTimestampsAndDelete(invalidFile, forceDelete, false);
            }
            // do the same for the index files.
            CarbonFile[] invalidIndexFiles = updateStatusManager.getUpdateDeltaFilesList(segment.getLoadName(), false, CarbonCommonConstants.UPDATE_INDEX_FILE_EXT, true, allSegmentFiles);
            for (CarbonFile invalidFile : invalidIndexFiles) {
                compareTimestampsAndDelete(invalidFile, forceDelete, false);
            }
            // now handle all the delete delta files which needs to be deleted.
            // there are 2 cases here .
            // 1. if the block is marked as compacted then the corresponding delta files
            //    can be deleted if query exec timeout is done.
            // 2. if the block is in success state then also there can be delete
            //    delta compaction happened and old files can be deleted.
            SegmentUpdateDetails[] updateDetails = updateStatusManager.readLoadMetadata();
            for (SegmentUpdateDetails block : updateDetails) {
                CarbonFile[] completeListOfDeleteDeltaFiles;
                CarbonFile[] invalidDeleteDeltaFiles;
                if (!block.getSegmentName().equalsIgnoreCase(segment.getLoadName())) {
                    continue;
                }
                // case 1
                if (CarbonUpdateUtil.isBlockInvalid(block.getStatus())) {
                    completeListOfDeleteDeltaFiles = updateStatusManager.getDeleteDeltaInvalidFilesList(segment.getLoadName(), block, true, allSegmentFiles);
                    for (CarbonFile invalidFile : completeListOfDeleteDeltaFiles) {
                        compareTimestampsAndDelete(invalidFile, forceDelete, false);
                    }
                    CarbonFile[] blockRelatedFiles = updateStatusManager.getAllBlockRelatedFiles(block.getBlockName(), allSegmentFiles, block.getActualBlockName());
                    for (CarbonFile invalidFile : blockRelatedFiles) {
                        compareTimestampsAndDelete(invalidFile, forceDelete, false);
                    }
                } else {
                    invalidDeleteDeltaFiles = updateStatusManager.getDeleteDeltaInvalidFilesList(segment.getLoadName(), block, false, allSegmentFiles);
                    for (CarbonFile invalidFile : invalidDeleteDeltaFiles) {
                        compareTimestampsAndDelete(invalidFile, forceDelete, false);
                    }
                }
            }
        }
    }
    // delete the update table status files which are old.
    if (null != validUpdateStatusFile && !validUpdateStatusFile.isEmpty()) {
        final String updateStatusTimestamp = validUpdateStatusFile.substring(validUpdateStatusFile.lastIndexOf(CarbonCommonConstants.HYPHEN) + 1);
        CarbonFile metaFolder = FileFactory.getCarbonFile(carbonTablePath.getMetadataDirectoryPath(), FileFactory.getFileType(carbonTablePath.getMetadataDirectoryPath()));
        CarbonFile[] invalidUpdateStatusFiles = metaFolder.listFiles(new CarbonFileFilter() {

            @Override
            public boolean accept(CarbonFile file) {
                if (file.getName().startsWith(CarbonCommonConstants.TABLEUPDATESTATUS_FILENAME)) {
                    // we only send invalid ones to delete.
                    if (!file.getName().endsWith(updateStatusTimestamp)) {
                        return true;
                    }
                }
                return false;
            }
        });
        for (CarbonFile invalidFile : invalidUpdateStatusFiles) {
            compareTimestampsAndDelete(invalidFile, forceDelete, true);
        }
    }
}

Example 2 with SegmentStatusManager

use of org.apache.carbondata.core.statusmanager.SegmentStatusManager in project carbondata by apache.

the class CarbonInputFormat method getSplits.

/**
   * {@inheritDoc}
   * Configurations FileInputFormat.INPUT_DIR
   * are used to get table path to read.
   *
   * @param job
   * @return List<InputSplit> list of CarbonInputSplit
   * @throws IOException
   */
@Override
public List<InputSplit> getSplits(JobContext job) throws IOException {
    AbsoluteTableIdentifier identifier = getAbsoluteTableIdentifier(job.getConfiguration());
    CacheClient cacheClient = new CacheClient(identifier.getStorePath());
    try {
        List<String> invalidSegments = new ArrayList<>();
        List<UpdateVO> invalidTimestampsList = new ArrayList<>();
        // get all valid segments and set them into the configuration
        if (getSegmentsToAccess(job).length == 0) {
            SegmentStatusManager segmentStatusManager = new SegmentStatusManager(identifier);
            SegmentStatusManager.ValidAndInvalidSegmentsInfo segments = segmentStatusManager.getValidAndInvalidSegments();
            SegmentUpdateStatusManager updateStatusManager = new SegmentUpdateStatusManager(identifier);
            setSegmentsToAccess(job.getConfiguration(), segments.getValidSegments());
            if (segments.getValidSegments().size() == 0) {
                return new ArrayList<>(0);
            }
            // remove entry in the segment index if there are invalid segments
            invalidSegments.addAll(segments.getInvalidSegments());
            for (String invalidSegmentId : invalidSegments) {
                invalidTimestampsList.add(updateStatusManager.getInvalidTimestampRange(invalidSegmentId));
            }
            if (invalidSegments.size() > 0) {
                List<TableSegmentUniqueIdentifier> invalidSegmentsIds = new ArrayList<>(invalidSegments.size());
                for (String segId : invalidSegments) {
                    invalidSegmentsIds.add(new TableSegmentUniqueIdentifier(identifier, segId));
                }
                cacheClient.getSegmentAccessClient().invalidateAll(invalidSegmentsIds);
            }
        }
        // process and resolve the expression
        Expression filter = getFilterPredicates(job.getConfiguration());
        CarbonTable carbonTable = getCarbonTable(job.getConfiguration());
        // this will be null in case of corrupt schema file.
        if (null == carbonTable) {
            throw new IOException("Missing/Corrupt schema file for table.");
        }
        CarbonInputFormatUtil.processFilterExpression(filter, carbonTable);
        // prune partitions for filter query on partition table
        BitSet matchedPartitions = null;
        if (null != filter) {
            PartitionInfo partitionInfo = carbonTable.getPartitionInfo(carbonTable.getFactTableName());
            if (null != partitionInfo) {
                Partitioner partitioner = PartitionUtil.getPartitioner(partitionInfo);
                matchedPartitions = new FilterExpressionProcessor().getFilteredPartitions(filter, partitionInfo, partitioner);
                if (matchedPartitions.cardinality() == 0) {
                    // no partition is required
                    return new ArrayList<InputSplit>();
                }
                if (matchedPartitions.cardinality() == partitioner.numPartitions()) {
                    // all partitions are required, no need to prune partitions
                    matchedPartitions = null;
                }
            }
        }
        FilterResolverIntf filterInterface = CarbonInputFormatUtil.resolveFilter(filter, identifier);
        // do block filtering and get split
        List<InputSplit> splits = getSplits(job, filterInterface, matchedPartitions, cacheClient);
        // pass the invalid segment to task side in order to remove index entry in task side
        if (invalidSegments.size() > 0) {
            for (InputSplit split : splits) {
                ((CarbonInputSplit) split).setInvalidSegments(invalidSegments);
                ((CarbonInputSplit) split).setInvalidTimestampRange(invalidTimestampsList);
            }
        }
        return splits;
    } finally {
        // close the cache cache client to clear LRU cache memory
        cacheClient.close();
    }
}

Example 3 with SegmentStatusManager

use of org.apache.carbondata.core.statusmanager.SegmentStatusManager in project carbondata by apache.

the class CarbonDataMergerUtil method updateMajorCompactionPropertyInSegment.

/**
   * This will update the property of segments as major compacted.
   * @param model
   * @param changedSegDetails
   */
public static void updateMajorCompactionPropertyInSegment(CarbonLoadModel model, List<LoadMetadataDetails> changedSegDetails, List<LoadMetadataDetails> preservedSegment) throws Exception {
    String metadataPath = model.getCarbonDataLoadSchema().getCarbonTable().getMetaDataFilepath();
    AbsoluteTableIdentifier absoluteTableIdentifier = model.getCarbonDataLoadSchema().getCarbonTable().getAbsoluteTableIdentifier();
    SegmentStatusManager segmentStatusManager = new SegmentStatusManager(absoluteTableIdentifier);
    LoadMetadataDetails[] details = segmentStatusManager.readLoadMetadata(metadataPath);
    List<LoadMetadataDetails> originalList = Arrays.asList(details);
    for (LoadMetadataDetails segment : changedSegDetails) {
        if (preservedSegment.contains(segment)) {
            continue;
        }
        originalList.get(originalList.indexOf(segment)).setMajorCompacted("true");
    }
    ICarbonLock carbonTableStatusLock = CarbonLockFactory.getCarbonLockObj(model.getCarbonDataLoadSchema().getCarbonTable().getCarbonTableIdentifier(), LockUsage.TABLE_STATUS_LOCK);
    try {
        if (carbonTableStatusLock.lockWithRetries()) {
            LOGGER.info("Acquired lock for the table " + model.getDatabaseName() + "." + model.getTableName() + " for table status updation ");
            CarbonTablePath carbonTablePath = CarbonStorePath.getCarbonTablePath(absoluteTableIdentifier.getStorePath(), absoluteTableIdentifier.getCarbonTableIdentifier());
            segmentStatusManager.writeLoadDetailsIntoFile(carbonTablePath.getTableStatusFilePath(), originalList.toArray(new LoadMetadataDetails[originalList.size()]));
        } else {
            LOGGER.error("Could not able to obtain lock for table" + model.getDatabaseName() + "." + model.getTableName() + "for table status updation");
            throw new Exception("Failed to update the MajorCompactionStatus.");
        }
    } catch (IOException e) {
        LOGGER.error("Error while writing metadata");
        throw new Exception("Failed to update the MajorCompactionStatus." + e.getMessage());
    } finally {
        if (carbonTableStatusLock.unlock()) {
            LOGGER.info("Table unlocked successfully after table status updation" + model.getDatabaseName() + "." + model.getTableName());
        } else {
            LOGGER.error("Unable to unlock Table lock for table" + model.getDatabaseName() + "." + model.getTableName() + " during table status updation");
        }
    }
}

Example 4 with SegmentStatusManager

use of org.apache.carbondata.core.statusmanager.SegmentStatusManager in project carbondata by apache.

the class CarbonDataMergerUtil method updateLoadMetadataIUDUpdateDeltaMergeStatus.

/**
   * Update Both Segment Update Status and Table Status for the case of IUD Delete
   * delta compaction.
   *
   * @param loadsToMerge
   * @param metaDataFilepath
   * @param carbonLoadModel
   * @return
   */
public static boolean updateLoadMetadataIUDUpdateDeltaMergeStatus(List<LoadMetadataDetails> loadsToMerge, String metaDataFilepath, CarbonLoadModel carbonLoadModel) {
    boolean status = false;
    boolean updateLockStatus = false;
    boolean tableLockStatus = false;
    String timestamp = "" + carbonLoadModel.getFactTimeStamp();
    List<String> updatedDeltaFilesList = new ArrayList<>(CarbonCommonConstants.DEFAULT_COLLECTION_SIZE);
    // This routine updateLoadMetadataIUDCompactionMergeStatus is suppose to update
    // two files as it is only called during IUD_UPDDEL_DELTA_COMPACTION. Along with
    // Table Status Metadata file (For Update Block Compaction) it has to update the
    // Table Update Status Metadata File (For corresponding Delete Delta File).
    // As the IUD_UPDDEL_DELTA_COMPACTION going to write in the same segment therefore in
    // A) Table Update Status Metadata File (Block Level)
    //      * For each blocks which is being compacted Mark 'Compacted' as the Status.
    // B) Table Status Metadata file (Segment Level)
    //      * loadStatus won't be changed to "compacted'
    //      * UpdateDeltaStartTime and UpdateDeltaEndTime will be both set to current
    //        timestamp (which is being passed from driver)
    // First the Table Update Status Metadata File should be updated as we need to get
    // the updated blocks for the segment from Table Status Metadata Update Delta Start and
    // End Timestamp.
    // Table Update Status Metadata Update.
    AbsoluteTableIdentifier absoluteTableIdentifier = carbonLoadModel.getCarbonDataLoadSchema().getCarbonTable().getAbsoluteTableIdentifier();
    CarbonTablePath carbonTablePath = CarbonStorePath.getCarbonTablePath(absoluteTableIdentifier.getStorePath(), absoluteTableIdentifier.getCarbonTableIdentifier());
    SegmentUpdateStatusManager segmentUpdateStatusManager = new SegmentUpdateStatusManager(absoluteTableIdentifier);
    SegmentStatusManager segmentStatusManager = new SegmentStatusManager(absoluteTableIdentifier);
    ICarbonLock updateLock = segmentUpdateStatusManager.getTableUpdateStatusLock();
    ICarbonLock statusLock = segmentStatusManager.getTableStatusLock();
    // Update the Compacted Blocks with Compacted Status.
    try {
        updatedDeltaFilesList = segmentUpdateStatusManager.getUpdateDeltaFiles(loadsToMerge.get(0).getLoadName().toString());
    } catch (Exception e) {
        LOGGER.error("Error while getting the Update Delta Blocks.");
        status = false;
        return status;
    }
    if (updatedDeltaFilesList.size() > 0) {
        try {
            updateLockStatus = updateLock.lockWithRetries();
            tableLockStatus = statusLock.lockWithRetries();
            List<String> blockNames = new ArrayList<>(updatedDeltaFilesList.size());
            for (String compactedBlocks : updatedDeltaFilesList) {
                // Try to BlockName
                int endIndex = compactedBlocks.lastIndexOf(File.separator);
                String blkNoExt = compactedBlocks.substring(endIndex + 1, compactedBlocks.lastIndexOf("-"));
                blockNames.add(blkNoExt);
            }
            if (updateLockStatus && tableLockStatus) {
                SegmentUpdateDetails[] updateLists = segmentUpdateStatusManager.readLoadMetadata();
                for (String compactedBlocks : blockNames) {
                    // Check is the compactedBlocks name matches with oldDetails
                    for (int i = 0; i < updateLists.length; i++) {
                        if (updateLists[i].getBlockName().equalsIgnoreCase(compactedBlocks) && !CarbonCommonConstants.COMPACTED.equalsIgnoreCase(updateLists[i].getStatus()) && !CarbonCommonConstants.MARKED_FOR_DELETE.equalsIgnoreCase(updateLists[i].getStatus())) {
                            updateLists[i].setStatus(CarbonCommonConstants.COMPACTED);
                        }
                    }
                }
                LoadMetadataDetails[] loadDetails = segmentStatusManager.readLoadMetadata(metaDataFilepath);
                for (LoadMetadataDetails loadDetail : loadDetails) {
                    if (loadsToMerge.contains(loadDetail)) {
                        loadDetail.setUpdateDeltaStartTimestamp(timestamp);
                        loadDetail.setUpdateDeltaEndTimestamp(timestamp);
                        if (loadDetail.getLoadName().equalsIgnoreCase("0")) {
                            loadDetail.setUpdateStatusFileName(CarbonUpdateUtil.getUpdateStatusFileName(timestamp));
                        }
                    }
                }
                try {
                    segmentUpdateStatusManager.writeLoadDetailsIntoFile(Arrays.asList(updateLists), timestamp);
                    segmentStatusManager.writeLoadDetailsIntoFile(carbonTablePath.getTableStatusFilePath(), loadDetails);
                    status = true;
                } catch (IOException e) {
                    LOGGER.error("Error while writing metadata. The metadata file path is " + carbonTablePath.getMetadataDirectoryPath());
                    status = false;
                }
            } else {
                LOGGER.error("Not able to acquire the lock.");
                status = false;
            }
        } catch (Exception e) {
            LOGGER.error("Error while updating metadata. The metadata file path is " + carbonTablePath.getMetadataDirectoryPath());
            status = false;
        } finally {
            if (updateLockStatus) {
                if (updateLock.unlock()) {
                    LOGGER.info("Unlock the segment update lock successfully.");
                } else {
                    LOGGER.error("Not able to unlock the segment update lock.");
                }
            }
            if (tableLockStatus) {
                if (statusLock.unlock()) {
                    LOGGER.info("Unlock the table status lock successfully.");
                } else {
                    LOGGER.error("Not able to unlock the table status lock.");
                }
            }
        }
    }
    return status;
}

Example 5 with SegmentStatusManager

use of org.apache.carbondata.core.statusmanager.SegmentStatusManager in project carbondata by apache.

the class CarbonDataMergerUtil method updateLoadMetadataWithMergeStatus.

/**
   * method to update table status in case of IUD Update Delta Compaction.
   * @param loadsToMerge
   * @param metaDataFilepath
   * @param MergedLoadName
   * @param carbonLoadModel
   * @param compactionType
   * @return
   */
public static boolean updateLoadMetadataWithMergeStatus(List<LoadMetadataDetails> loadsToMerge, String metaDataFilepath, String MergedLoadName, CarbonLoadModel carbonLoadModel, long mergeLoadStartTime, CompactionType compactionType) {
    boolean tableStatusUpdationStatus = false;
    AbsoluteTableIdentifier absoluteTableIdentifier = carbonLoadModel.getCarbonDataLoadSchema().getCarbonTable().getAbsoluteTableIdentifier();
    SegmentStatusManager segmentStatusManager = new SegmentStatusManager(absoluteTableIdentifier);
    ICarbonLock carbonLock = segmentStatusManager.getTableStatusLock();
    try {
        if (carbonLock.lockWithRetries()) {
            LOGGER.info("Acquired lock for the table " + carbonLoadModel.getDatabaseName() + "." + carbonLoadModel.getTableName() + " for table status updation ");
            CarbonTablePath carbonTablePath = CarbonStorePath.getCarbonTablePath(absoluteTableIdentifier.getStorePath(), absoluteTableIdentifier.getCarbonTableIdentifier());
            String statusFilePath = carbonTablePath.getTableStatusFilePath();
            LoadMetadataDetails[] loadDetails = SegmentStatusManager.readLoadMetadata(metaDataFilepath);
            String mergedLoadNumber = MergedLoadName.substring(MergedLoadName.lastIndexOf(CarbonCommonConstants.LOAD_FOLDER) + CarbonCommonConstants.LOAD_FOLDER.length(), MergedLoadName.length());
            long modificationOrDeletionTimeStamp = CarbonUpdateUtil.readCurrentTime();
            for (LoadMetadataDetails loadDetail : loadDetails) {
                // check if this segment is merged.
                if (loadsToMerge.contains(loadDetail)) {
                    // then need to discard the compaction process and treat it as failed compaction.
                    if (loadDetail.getLoadStatus().equalsIgnoreCase(CarbonCommonConstants.MARKED_FOR_DELETE)) {
                        LOGGER.error("Compaction is aborted as the segment " + loadDetail.getLoadName() + " is deleted after the compaction is started.");
                        return false;
                    }
                    loadDetail.setLoadStatus(CarbonCommonConstants.COMPACTED);
                    loadDetail.setModificationOrdeletionTimesStamp(modificationOrDeletionTimeStamp);
                    loadDetail.setMergedLoadName(mergedLoadNumber);
                }
            }
            // create entry for merged one.
            LoadMetadataDetails loadMetadataDetails = new LoadMetadataDetails();
            loadMetadataDetails.setPartitionCount(carbonLoadModel.getPartitionId());
            loadMetadataDetails.setLoadStatus(CarbonCommonConstants.STORE_LOADSTATUS_SUCCESS);
            long loadEnddate = CarbonUpdateUtil.readCurrentTime();
            loadMetadataDetails.setLoadEndTime(loadEnddate);
            loadMetadataDetails.setLoadName(mergedLoadNumber);
            loadMetadataDetails.setLoadStartTime(mergeLoadStartTime);
            loadMetadataDetails.setPartitionCount("0");
            // if this is a major compaction then set the segment as major compaction.
            if (compactionType == CompactionType.MAJOR_COMPACTION) {
                loadMetadataDetails.setMajorCompacted("true");
            }
            List<LoadMetadataDetails> updatedDetailsList = new ArrayList<>(Arrays.asList(loadDetails));
            // put the merged folder entry
            updatedDetailsList.add(loadMetadataDetails);
            try {
                SegmentStatusManager.writeLoadDetailsIntoFile(statusFilePath, updatedDetailsList.toArray(new LoadMetadataDetails[updatedDetailsList.size()]));
                tableStatusUpdationStatus = true;
            } catch (IOException e) {
                LOGGER.error("Error while writing metadata");
                tableStatusUpdationStatus = false;
            }
        } else {
            LOGGER.error("Could not able to obtain lock for table" + carbonLoadModel.getDatabaseName() + "." + carbonLoadModel.getTableName() + "for table status updation");
        }
    } finally {
        if (carbonLock.unlock()) {
            LOGGER.info("Table unlocked successfully after table status updation" + carbonLoadModel.getDatabaseName() + "." + carbonLoadModel.getTableName());
        } else {
            LOGGER.error("Unable to unlock Table lock for table" + carbonLoadModel.getDatabaseName() + "." + carbonLoadModel.getTableName() + " during table status updation");
        }
    }
    return tableStatusUpdationStatus;
}