apache/pinot

Refactor to centralize Job Specification properties for reusability across ingestion jobs

Open

#14,272 建立於 2024年10月22日

在 GitHub 查看
 (0 留言) (0 反應) (0 負責人)Java (4,937 star) (1,234 fork)batch import
cleanupgood first issueingestion

描述

Currently some of the the job specification properties are defined at each implementation of batch ingestion like

Hadoop job runner has defined the job spec constants in it's corresponding implementation:

public static final String SEGMENT_GENERATION_JOB_SPEC = "segmentGenerationJobSpec";

// Field names in job spec's executionFrameworkSpec/extraConfigs section private static final String DEPS_JAR_DIR_FIELD = "dependencyJarDir"; private static final String STAGING_DIR_FIELD = "stagingDir";

Similarly spark ingestion runner also defined the same constants in its own classes

private static final Logger LOGGER = LoggerFactory.getLogger(SparkSegmentGenerationJobRunner.class); private static final String DEPS_JAR_DIR = "dependencyJarDir"; private static final String STAGING_DIR = "stagingDir";

Better to centralize them at common place and reuse to avoid deviations in the future.

貢獻者指南

Refactor to centralize Job Specification properties for reusability across ingestion jobs · apache/pinot#14272 | Good First Issue