swaminathanmanish commented on code in PR #14300: URL: https://github.com/apache/pinot/pull/14300#discussion_r1842031808
########## pinot-plugins/pinot-minion-tasks/pinot-minion-builtin-tasks/src/main/java/org/apache/pinot/plugin/minion/tasks/refreshsegment/RefreshSegmentTaskExecutor.java: ########## @@ -0,0 +1,203 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.pinot.plugin.minion.tasks.refreshsegment; + +import java.io.File; +import java.util.Collections; +import java.util.HashSet; +import java.util.Map; +import java.util.Set; +import org.apache.pinot.common.metadata.segment.SegmentZKMetadataCustomMapModifier; +import org.apache.pinot.core.common.MinionConstants; +import org.apache.pinot.core.minion.PinotTaskConfig; +import org.apache.pinot.plugin.minion.tasks.BaseSingleSegmentConversionExecutor; +import org.apache.pinot.plugin.minion.tasks.SegmentConversionResult; +import org.apache.pinot.segment.local.indexsegment.immutable.ImmutableSegmentLoader; +import org.apache.pinot.segment.local.segment.creator.impl.SegmentIndexCreationDriverImpl; +import org.apache.pinot.segment.local.segment.index.loader.IndexLoadingConfig; +import org.apache.pinot.segment.local.segment.readers.PinotSegmentRecordReader; +import org.apache.pinot.segment.spi.ColumnMetadata; +import org.apache.pinot.segment.spi.creator.SegmentGeneratorConfig; +import org.apache.pinot.segment.spi.index.metadata.SegmentMetadataImpl; +import org.apache.pinot.segment.spi.loader.SegmentDirectoryLoaderContext; +import org.apache.pinot.segment.spi.loader.SegmentDirectoryLoaderRegistry; +import org.apache.pinot.segment.spi.store.SegmentDirectory; +import org.apache.pinot.spi.config.table.TableConfig; +import org.apache.pinot.spi.data.FieldSpec; +import org.apache.pinot.spi.data.Schema; +import org.apache.pinot.spi.env.PinotConfiguration; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + + +public class RefreshSegmentTaskExecutor extends BaseSingleSegmentConversionExecutor { + private static final Logger LOGGER = LoggerFactory.getLogger(RefreshSegmentTaskGenerator.class); + + private long _taskStartTime; + + /** + * The code here currently covers segment refresh for the following cases: + * 1. Process newly added columns. + * 2. Addition/removal of indexes. + * 3. Compatible datatype change for existing columns + */ + @Override + protected SegmentConversionResult convert(PinotTaskConfig pinotTaskConfig, File indexDir, File workingDir) + throws Exception { + _eventObserver.notifyProgress(pinotTaskConfig, "Refreshing segment: " + indexDir); + + // We set _taskStartTime before fetching the tableConfig. Task Generation relies on tableConfig/Schema updates + // happening after the last processed time. So we explicity use the timestamp before fetching tableConfig as the + // processedTime. + _taskStartTime = System.currentTimeMillis(); + Map<String, String> configs = pinotTaskConfig.getConfigs(); + String tableNameWithType = configs.get(MinionConstants.TABLE_NAME_KEY); + String segmentName = configs.get(MinionConstants.SEGMENT_NAME_KEY); + String taskType = pinotTaskConfig.getTaskType(); + + LOGGER.info("Starting task: {} with configs: {}", taskType, configs); + + TableConfig tableConfig = getTableConfig(tableNameWithType); + Schema schema = getSchema(tableNameWithType); + + IndexLoadingConfig indexLoadingConfig = new IndexLoadingConfig(tableConfig, schema); + SegmentMetadataImpl segmentMetadata = new SegmentMetadataImpl(indexDir); + PinotConfiguration segmentDirectoryConfigs = indexLoadingConfig.getSegmentDirectoryConfigs(); + SegmentDirectoryLoaderContext segmentLoaderContext = + new SegmentDirectoryLoaderContext.Builder().setTableConfig(indexLoadingConfig.getTableConfig()) + .setSchema(schema) + .setInstanceId(indexLoadingConfig.getInstanceId()) + .setSegmentName(segmentMetadata.getName()) + .setSegmentCrc(segmentMetadata.getCrc()) + .setSegmentDirectoryConfigs(segmentDirectoryConfigs) + .build(); + SegmentDirectory segmentDirectory = + SegmentDirectoryLoaderRegistry.getDefaultSegmentDirectoryLoader().load(indexDir.toURI(), segmentLoaderContext); + + // BaseDefaultColumnHandler part of needPreprocess() does not process any changes to existing columns like datatype, + // change from dimension to metric, etc. + boolean needPreprocess = ImmutableSegmentLoader.needPreprocess(segmentDirectory, indexLoadingConfig, schema); + closeSegmentDirectoryQuietly(segmentDirectory); + Set<String> refreshColumnSet = new HashSet<>(); + + for (FieldSpec fieldSpecInSchema : schema.getAllFieldSpecs()) { + // Virtual columns are constructed while loading the segment, thus do not exist in the record, nor should be + // persisted to the disk. + if (fieldSpecInSchema.isVirtualColumn()) { + continue; + } + + String column = fieldSpecInSchema.getName(); + ColumnMetadata columnMetadata = segmentMetadata.getColumnMetadataFor(column); + if (columnMetadata != null) { + FieldSpec fieldSpecInSegment = columnMetadata.getFieldSpec(); + + // Check the data type and default value matches. + FieldSpec.DataType dataTypeInSegment = fieldSpecInSegment.getDataType(); + FieldSpec.DataType dataTypeInSchema = fieldSpecInSchema.getDataType(); + + // Column exists in segment. + if (dataTypeInSegment != dataTypeInSchema) { + // Check if we need to update the data-type. DataType change is dependent on segmentGeneration code converting + // the object to the destination datatype. If the existing data is the column is not compatible with the + // destination data-type, the refresh task will fail. + refreshColumnSet.add(column); + } + + // TODO: Maybe we can support singleValue to multi-value conversions are supproted and vice-versa. + } else { + refreshColumnSet.add(column); + } + } + + if (!needPreprocess && refreshColumnSet.isEmpty()) { Review Comment: This is the no-op case where we dont need to do anything to that segment, right? In **BaseSingleSegmentConversionExecutor** looks like the upload is skipped if the file is missing ``` File convertedSegmentDir = segmentConversionResult.getFile(); if (convertedSegmentDir == null) { return segmentConversionResult; } ``` Above means we will not update Segment Zk metadata as well which comes later in that function. If thats the case, would we be picking the same segments again and again, when task generator is triggered (i.e when refresh is not needed) ? ``` if (!shouldRefreshSegment(segmentZKMetadata, tableConfig, tableStat, schemaStat)) { continue; } ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
