A Job is the top-level entity representing an entire batch process, composed of one or more Steps executed in sequence (or conditionally). Each Step encapsulates one independent phase of processing (e.g., 'import file', 'validate records', 'generate summary report') and does the actual work. The JobRepository is the persistence layer tracking metadata about every JobInstance, JobExecution, and StepExecution — including status, start/end times, and read/write counts — which is exactly what makes restart and monitoring possible.

Intermediate5 min

AccentureCapgeminiCognizantOracleIBMAmazon

#Job#Step#JobRepository

What is the relationship between a Job, a Step, and a JobRepository in Spring Batch?

JobRepository Is What Makes 'Did This Already Run?' a Solvable Question

Because JobRepository persistently records every execution attempt with its parameters, Spring Batch can definitively answer whether a job with a given set of parameters already completed successfully — attempting to re-launch an already-completed JobInstance with identical parameters throws a JobInstanceAlreadyCompleteException by default, which is a deliberate safeguard against accidentally reprocessing the same data twice, not a bug to work around.

A Simple Two-Step Job

java

@Bean
Job importJob(JobRepository repo, Step importStep, Step reportStep) {
    return new JobBuilder("importJob", repo)
        .start(importStep)
        .next(reportStep)
        .build();
}