Skip to content

Commit

Permalink
Remove TransmissionJobOption.buildTaskConfiguration()
Browse files Browse the repository at this point in the history
  • Loading branch information
terrymanu committed Dec 1, 2023
1 parent 28814e9 commit f070d1d
Show file tree
Hide file tree
Showing 6 changed files with 164 additions and 134 deletions.
Original file line number Diff line number Diff line change
Expand Up @@ -18,13 +18,11 @@
package org.apache.shardingsphere.data.pipeline.core.job.option;

import org.apache.shardingsphere.data.pipeline.common.config.job.PipelineJobConfiguration;
import org.apache.shardingsphere.data.pipeline.common.config.process.PipelineProcessConfiguration;
import org.apache.shardingsphere.data.pipeline.common.context.TransmissionProcessContext;
import org.apache.shardingsphere.data.pipeline.common.job.progress.yaml.YamlTransmissionJobItemProgressSwapper;
import org.apache.shardingsphere.data.pipeline.common.pojo.PipelineJobInfo;
import org.apache.shardingsphere.data.pipeline.core.consistencycheck.ConsistencyCheckJobItemProgressContext;
import org.apache.shardingsphere.data.pipeline.core.consistencycheck.PipelineDataConsistencyChecker;
import org.apache.shardingsphere.data.pipeline.core.task.config.PipelineTaskConfiguration;

/**
* Transmission job option.
Expand All @@ -45,16 +43,6 @@ default YamlTransmissionJobItemProgressSwapper getYamlJobItemProgressSwapper() {
*/
PipelineJobInfo getJobInfo(String jobId);

/**
* Build task configuration.
*
* @param jobConfig pipeline job configuration
* @param jobShardingItem job sharding item
* @param processConfig pipeline process configuration
* @return task configuration
*/
PipelineTaskConfiguration buildTaskConfiguration(PipelineJobConfiguration jobConfig, int jobShardingItem, PipelineProcessConfiguration processConfig);

/**
* Build transmission process context.
*
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -20,25 +20,38 @@
import lombok.Getter;
import lombok.RequiredArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import org.apache.shardingsphere.data.pipeline.api.PipelineDataSourceConfiguration;
import org.apache.shardingsphere.data.pipeline.api.type.StandardPipelineDataSourceConfiguration;
import org.apache.shardingsphere.data.pipeline.cdc.api.CDCJobAPI;
import org.apache.shardingsphere.data.pipeline.cdc.config.job.CDCJobConfiguration;
import org.apache.shardingsphere.data.pipeline.cdc.config.task.CDCTaskConfiguration;
import org.apache.shardingsphere.data.pipeline.cdc.config.yaml.YamlCDCJobConfigurationSwapper;
import org.apache.shardingsphere.data.pipeline.cdc.context.CDCJobItemContext;
import org.apache.shardingsphere.data.pipeline.cdc.context.CDCProcessContext;
import org.apache.shardingsphere.data.pipeline.cdc.core.importer.sink.CDCSocketSink;
import org.apache.shardingsphere.data.pipeline.cdc.core.prepare.CDCJobPreparer;
import org.apache.shardingsphere.data.pipeline.cdc.core.task.CDCTasksRunner;
import org.apache.shardingsphere.data.pipeline.cdc.generator.CDCResponseUtils;
import org.apache.shardingsphere.data.pipeline.cdc.config.yaml.YamlCDCJobConfigurationSwapper;
import org.apache.shardingsphere.data.pipeline.common.config.ImporterConfiguration;
import org.apache.shardingsphere.data.pipeline.common.config.process.PipelineProcessConfiguration;
import org.apache.shardingsphere.data.pipeline.common.context.PipelineJobItemContext;
import org.apache.shardingsphere.data.pipeline.common.datanode.JobDataNodeLine;
import org.apache.shardingsphere.data.pipeline.common.datanode.JobDataNodeLineConvertUtils;
import org.apache.shardingsphere.data.pipeline.common.datasource.DefaultPipelineDataSourceManager;
import org.apache.shardingsphere.data.pipeline.common.datasource.PipelineDataSourceConfigurationFactory;
import org.apache.shardingsphere.data.pipeline.common.datasource.PipelineDataSourceManager;
import org.apache.shardingsphere.data.pipeline.common.execute.ExecuteCallback;
import org.apache.shardingsphere.data.pipeline.common.execute.ExecuteEngine;
import org.apache.shardingsphere.data.pipeline.common.ingest.position.FinishedPosition;
import org.apache.shardingsphere.data.pipeline.common.job.JobStatus;
import org.apache.shardingsphere.data.pipeline.common.job.progress.TransmissionJobItemProgress;
import org.apache.shardingsphere.data.pipeline.common.metadata.CaseInsensitiveIdentifier;
import org.apache.shardingsphere.data.pipeline.common.spi.algorithm.JobRateLimitAlgorithm;
import org.apache.shardingsphere.data.pipeline.common.util.ShardingColumnsExtractor;
import org.apache.shardingsphere.data.pipeline.core.importer.sink.PipelineSink;
import org.apache.shardingsphere.data.pipeline.core.ingest.dumper.context.DumperCommonContext;
import org.apache.shardingsphere.data.pipeline.core.ingest.dumper.context.IncrementalDumperContext;
import org.apache.shardingsphere.data.pipeline.core.ingest.dumper.context.mapper.TableAndSchemaNameMapper;
import org.apache.shardingsphere.data.pipeline.core.job.AbstractPipelineJob;
import org.apache.shardingsphere.data.pipeline.core.job.PipelineJobCenter;
import org.apache.shardingsphere.data.pipeline.core.job.PipelineJobIdUtils;
Expand All @@ -55,8 +68,11 @@
import java.util.Collection;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Optional;
import java.util.Set;
import java.util.concurrent.CompletableFuture;
import java.util.stream.Collectors;

/**
* CDC job.
Expand Down Expand Up @@ -93,7 +109,7 @@ public void execute(final ShardingContext shardingContext) {
log.info("stopping true, ignore");
return;
}
CDCJobItemContext jobItemContext = buildPipelineJobItemContext(jobConfig, shardingItem);
CDCJobItemContext jobItemContext = buildCDCJobItemContext(jobConfig, shardingItem);
PipelineTasksRunner tasksRunner = new CDCTasksRunner(jobItemContext);
if (!addTasksRunner(shardingItem, tasksRunner)) {
continue;
Expand All @@ -111,13 +127,43 @@ public void execute(final ShardingContext shardingContext) {
executeIncrementalTasks(jobItemContexts);
}

private CDCJobItemContext buildPipelineJobItemContext(final CDCJobConfiguration jobConfig, final int shardingItem) {
private CDCJobItemContext buildCDCJobItemContext(final CDCJobConfiguration jobConfig, final int shardingItem) {
Optional<TransmissionJobItemProgress> initProgress = jobItemManager.getProgress(jobConfig.getJobId(), shardingItem);
CDCProcessContext jobProcessContext = jobOption.buildProcessContext(jobConfig);
CDCTaskConfiguration taskConfig = jobOption.buildTaskConfiguration(jobConfig, shardingItem, jobProcessContext.getPipelineProcessConfig());
CDCTaskConfiguration taskConfig = buildTaskConfiguration(jobConfig, shardingItem, jobProcessContext.getPipelineProcessConfig());
return new CDCJobItemContext(jobConfig, shardingItem, initProgress.orElse(null), jobProcessContext, taskConfig, dataSourceManager, sink);
}

private CDCTaskConfiguration buildTaskConfiguration(final CDCJobConfiguration jobConfig, final int jobShardingItem, final PipelineProcessConfiguration processConfig) {
TableAndSchemaNameMapper tableAndSchemaNameMapper = new TableAndSchemaNameMapper(jobConfig.getSchemaTableNames());
IncrementalDumperContext dumperContext = buildDumperContext(jobConfig, jobShardingItem, tableAndSchemaNameMapper);
ImporterConfiguration importerConfig = buildImporterConfiguration(jobConfig, processConfig, jobConfig.getSchemaTableNames(), tableAndSchemaNameMapper);
CDCTaskConfiguration result = new CDCTaskConfiguration(dumperContext, importerConfig);
log.debug("buildTaskConfiguration, result={}", result);
return result;
}

private IncrementalDumperContext buildDumperContext(final CDCJobConfiguration jobConfig, final int jobShardingItem, final TableAndSchemaNameMapper tableAndSchemaNameMapper) {
JobDataNodeLine dataNodeLine = jobConfig.getJobShardingDataNodes().get(jobShardingItem);
String dataSourceName = dataNodeLine.getEntries().iterator().next().getDataNodes().iterator().next().getDataSourceName();
StandardPipelineDataSourceConfiguration actualDataSourceConfig = jobConfig.getDataSourceConfig().getActualDataSourceConfiguration(dataSourceName);
return new IncrementalDumperContext(
new DumperCommonContext(dataSourceName, actualDataSourceConfig, JobDataNodeLineConvertUtils.buildTableNameMapper(dataNodeLine), tableAndSchemaNameMapper),
jobConfig.getJobId(), jobConfig.isDecodeWithTX());
}

private ImporterConfiguration buildImporterConfiguration(final CDCJobConfiguration jobConfig, final PipelineProcessConfiguration pipelineProcessConfig, final Collection<String> schemaTableNames,
final TableAndSchemaNameMapper tableAndSchemaNameMapper) {
PipelineDataSourceConfiguration dataSourceConfig = PipelineDataSourceConfigurationFactory.newInstance(jobConfig.getDataSourceConfig().getType(),
jobConfig.getDataSourceConfig().getParameter());
CDCProcessContext processContext = new CDCProcessContext(jobConfig.getJobId(), pipelineProcessConfig);
JobRateLimitAlgorithm writeRateLimitAlgorithm = processContext.getWriteRateLimitAlgorithm();
int batchSize = pipelineProcessConfig.getWrite().getBatchSize();
Map<CaseInsensitiveIdentifier, Set<String>> shardingColumnsMap = new ShardingColumnsExtractor()
.getShardingColumnsMap(jobConfig.getDataSourceConfig().getRootConfig().getRules(), schemaTableNames.stream().map(CaseInsensitiveIdentifier::new).collect(Collectors.toSet()));
return new ImporterConfiguration(dataSourceConfig, shardingColumnsMap, tableAndSchemaNameMapper, batchSize, writeRateLimitAlgorithm, 0, 1);
}

private void prepare(final Collection<CDCJobItemContext> jobItemContexts) {
try {
jobPreparer.initTasks(jobItemContexts);
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -18,39 +18,20 @@
package org.apache.shardingsphere.data.pipeline.cdc;

import lombok.extern.slf4j.Slf4j;
import org.apache.shardingsphere.data.pipeline.api.PipelineDataSourceConfiguration;
import org.apache.shardingsphere.data.pipeline.api.type.StandardPipelineDataSourceConfiguration;
import org.apache.shardingsphere.data.pipeline.cdc.config.job.CDCJobConfiguration;
import org.apache.shardingsphere.data.pipeline.cdc.config.task.CDCTaskConfiguration;
import org.apache.shardingsphere.data.pipeline.cdc.config.yaml.YamlCDCJobConfigurationSwapper;
import org.apache.shardingsphere.data.pipeline.cdc.context.CDCProcessContext;
import org.apache.shardingsphere.data.pipeline.common.config.ImporterConfiguration;
import org.apache.shardingsphere.data.pipeline.common.config.job.PipelineJobConfiguration;
import org.apache.shardingsphere.data.pipeline.common.config.process.PipelineProcessConfiguration;
import org.apache.shardingsphere.data.pipeline.common.context.TransmissionProcessContext;
import org.apache.shardingsphere.data.pipeline.common.datanode.JobDataNodeLine;
import org.apache.shardingsphere.data.pipeline.common.datanode.JobDataNodeLineConvertUtils;
import org.apache.shardingsphere.data.pipeline.common.datasource.PipelineDataSourceConfigurationFactory;
import org.apache.shardingsphere.data.pipeline.common.metadata.CaseInsensitiveIdentifier;
import org.apache.shardingsphere.data.pipeline.common.pojo.PipelineJobInfo;
import org.apache.shardingsphere.data.pipeline.common.pojo.PipelineJobMetaData;
import org.apache.shardingsphere.data.pipeline.common.spi.algorithm.JobRateLimitAlgorithm;
import org.apache.shardingsphere.data.pipeline.common.util.ShardingColumnsExtractor;
import org.apache.shardingsphere.data.pipeline.core.consistencycheck.ConsistencyCheckJobItemProgressContext;
import org.apache.shardingsphere.data.pipeline.core.consistencycheck.PipelineDataConsistencyChecker;
import org.apache.shardingsphere.data.pipeline.core.ingest.dumper.context.DumperCommonContext;
import org.apache.shardingsphere.data.pipeline.core.ingest.dumper.context.IncrementalDumperContext;
import org.apache.shardingsphere.data.pipeline.core.ingest.dumper.context.mapper.TableAndSchemaNameMapper;
import org.apache.shardingsphere.data.pipeline.core.job.PipelineJobIdUtils;
import org.apache.shardingsphere.data.pipeline.core.job.option.TransmissionJobOption;
import org.apache.shardingsphere.data.pipeline.core.job.service.PipelineJobConfigurationManager;
import org.apache.shardingsphere.data.pipeline.core.job.service.TransmissionJobManager;

import java.util.Collection;
import java.util.Map;
import java.util.Set;
import java.util.stream.Collectors;

/**
* CDC job option.
*/
Expand Down Expand Up @@ -80,38 +61,6 @@ public PipelineJobInfo getJobInfo(final String jobId) {
return new PipelineJobInfo(jobMetaData, jobConfig.getDatabaseName(), String.join(", ", jobConfig.getSchemaTableNames()));
}

@Override
public CDCTaskConfiguration buildTaskConfiguration(final PipelineJobConfiguration pipelineJobConfig, final int jobShardingItem, final PipelineProcessConfiguration processConfig) {
CDCJobConfiguration jobConfig = (CDCJobConfiguration) pipelineJobConfig;
TableAndSchemaNameMapper tableAndSchemaNameMapper = new TableAndSchemaNameMapper(jobConfig.getSchemaTableNames());
IncrementalDumperContext dumperContext = buildDumperContext(jobConfig, jobShardingItem, tableAndSchemaNameMapper);
ImporterConfiguration importerConfig = buildImporterConfiguration(jobConfig, processConfig, jobConfig.getSchemaTableNames(), tableAndSchemaNameMapper);
CDCTaskConfiguration result = new CDCTaskConfiguration(dumperContext, importerConfig);
log.debug("buildTaskConfiguration, result={}", result);
return result;
}

private IncrementalDumperContext buildDumperContext(final CDCJobConfiguration jobConfig, final int jobShardingItem, final TableAndSchemaNameMapper tableAndSchemaNameMapper) {
JobDataNodeLine dataNodeLine = jobConfig.getJobShardingDataNodes().get(jobShardingItem);
String dataSourceName = dataNodeLine.getEntries().iterator().next().getDataNodes().iterator().next().getDataSourceName();
StandardPipelineDataSourceConfiguration actualDataSourceConfig = jobConfig.getDataSourceConfig().getActualDataSourceConfiguration(dataSourceName);
return new IncrementalDumperContext(
new DumperCommonContext(dataSourceName, actualDataSourceConfig, JobDataNodeLineConvertUtils.buildTableNameMapper(dataNodeLine), tableAndSchemaNameMapper),
jobConfig.getJobId(), jobConfig.isDecodeWithTX());
}

private ImporterConfiguration buildImporterConfiguration(final CDCJobConfiguration jobConfig, final PipelineProcessConfiguration pipelineProcessConfig, final Collection<String> schemaTableNames,
final TableAndSchemaNameMapper tableAndSchemaNameMapper) {
PipelineDataSourceConfiguration dataSourceConfig = PipelineDataSourceConfigurationFactory.newInstance(jobConfig.getDataSourceConfig().getType(),
jobConfig.getDataSourceConfig().getParameter());
CDCProcessContext processContext = new CDCProcessContext(jobConfig.getJobId(), pipelineProcessConfig);
JobRateLimitAlgorithm writeRateLimitAlgorithm = processContext.getWriteRateLimitAlgorithm();
int batchSize = pipelineProcessConfig.getWrite().getBatchSize();
Map<CaseInsensitiveIdentifier, Set<String>> shardingColumnsMap = new ShardingColumnsExtractor()
.getShardingColumnsMap(jobConfig.getDataSourceConfig().getRootConfig().getRules(), schemaTableNames.stream().map(CaseInsensitiveIdentifier::new).collect(Collectors.toSet()));
return new ImporterConfiguration(dataSourceConfig, shardingColumnsMap, tableAndSchemaNameMapper, batchSize, writeRateLimitAlgorithm, 0, 1);
}

@Override
public CDCProcessContext buildProcessContext(final PipelineJobConfiguration jobConfig) {
TransmissionJobManager jobManager = new TransmissionJobManager(this);
Expand Down
Loading

0 comments on commit f070d1d

Please sign in to comment.