apache/hadoop · error · IllegalArgumentException

JobStory not available for job {}

Error message

JobStory not available for job {}

What it means

Statistics.generateJobStats(job, jobdesc) computes the gridmix job sequence id from the job configuration and throws IllegalArgumentException if seq >= 0 but the JobStory (job descriptor from the trace) is null. It is an internal invariant: every submitted gridmix job with a valid sequence number must carry its original trace job. The sole caller is JobSubmitter's statistics thread, so the exception surfaces from the submitter after job submission.

Source

Thrown at hadoop-tools/hadoop-gridmix/src/main/java/org/apache/hadoop/mapred/gridmix/Statistics.java:109

        public JobClient run() throws IOException {
          return new JobClient(new JobConf(conf));
        }
      });

    this.jtPollingInterval = pollingInterval;
    maxJobCompletedInInterval = conf.getInt(
      MAX_JOBS_COMPLETED_IN_POLL_INTERVAL_KEY, 1);
    this.startFlag = startFlag;
  }

  /**
   * Generates a job stats.
   */
  public static JobStats generateJobStats(Job job, JobStory jobdesc) {
    int seq = GridmixJob.getJobSeqId(job);
    // bail out if job description is missing for a job to be simulated
    if (seq >= 0 && jobdesc == null) {
      throw new IllegalArgumentException("JobStory not available for job " 
                                         + job.getJobID());
    }
    
    int maps = -1;
    int reds = -1;
    if (jobdesc != null) {
      // Note that the ZombieJob will return a >= 0 value
      maps = jobdesc.getNumberMaps();
      reds = jobdesc.getNumberReduces();
    }
    return new JobStats(maps, reds, job);
  }

  private static void addToNumMapsSubmitted(int numMaps) {
    numMapsSubmitted += numMaps;
  }

  private static void addToNumReducesSubmitted(int numReduces) {

View on GitHub (pinned to 2add963021)

Solutions

1. Ensure your JobFactory/GridmixJob supplies a non-null JobStory for every job that carries a gridmix sequence id
2. If the job legitimately has no trace descriptor, clear/omit the job sequence id so seq < 0 (stats are then recorded with maps/reds = -1)
3. Regenerate the job trace from a clean job history so no submitted job maps to a null descriptor

// before recording stats for a submitted gridmix job
JobStory desc = job.getJobDesc();
if (desc == null && GridmixJob.getJobSeqId(job.getJob()) >= 0) {
  // invariant violated upstream; fix the job factory instead of calling generateJobStats
}

try {
  JobStats stats = Statistics.generateJobStats(job, jobdesc);
} catch (IllegalArgumentException e) {
  // job carried a gridmix seq id but no JobStory; repair the factory/trace wiring
}

When it happens

Trigger: A custom GridmixJob or JobFactory submits a job whose configuration carries the job-seq-id property but whose JobStory argument is null; replaying a trace where a submitted job has no corresponding trace job record (null desc from the factory).

Common situations: Extending gridmix (custom job factories/mappers) and not wiring the JobStory through; truncated or malformed trace files yielding null descriptions; version mismatches in gridmix extensions after upgrading Hadoop. Not triggerable by ordinary CLI configuration.

Related errors

• Failed to rename %s to %s, file already exists or not empty!
• Truncate is not supported by BaiduBosFileSystem
• user accessKey and secretAccessKey should not be null
• NoSuchUpload
• Key '" + key + "' does not exist in BOS

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/77ce118e888019cb. Report an issue: GitHub.