apache/hadoop · error · IOException

Unable to get log information for task: {oldTaskAttemptID}

Error message

Unable to get log information for task: {oldTaskAttemptID}

What it means

Thrown by ClientServiceDelegate.getLogFilePath when the job is in a terminal state (SUCCEEDED/FAILED/KILLED/ERROR) and a task attempt id was given, but the TaskAttemptReport obtained from the AM/history proxy has a null containerId or null nodeManagerHost. Without those fields the client cannot compute the NodeManager log address, so it raises IOException instead of returning bogus LogParams.

Source

Thrown at hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/main/java/org/apache/hadoop/mapred/ClientServiceDelegate.java:512

        recordFactory.newRecordInstance(GetJobReportRequest.class);
    request.setJobId(jobId);

    JobReport report =
        ((GetJobReportResponse) invoke("getJobReport",
            GetJobReportRequest.class, request)).getJobReport();
    if (EnumSet.of(JobState.SUCCEEDED, JobState.FAILED, JobState.KILLED,
        JobState.ERROR).contains(report.getJobState())) {
      if (oldTaskAttemptID != null) {
        GetTaskAttemptReportRequest taRequest =
            recordFactory.newRecordInstance(GetTaskAttemptReportRequest.class);
        taRequest.setTaskAttemptId(TypeConverter.toYarn(oldTaskAttemptID));
        TaskAttemptReport taReport =
            ((GetTaskAttemptReportResponse) invoke("getTaskAttemptReport",
                GetTaskAttemptReportRequest.class, taRequest))
                .getTaskAttemptReport();
        if (taReport.getContainerId() == null
            || taReport.getNodeManagerHost() == null) {
          throw new IOException("Unable to get log information for task: "
              + oldTaskAttemptID);
        }
        return new LogParams(
            taReport.getContainerId().toString(),
            taReport.getContainerId().getApplicationAttemptId()
                .getApplicationId().toString(),
            NodeId.newInstance(taReport.getNodeManagerHost(),
                taReport.getNodeManagerPort()).toString(), report.getUser());
      } else {
        if (report.getAMInfos() == null || report.getAMInfos().size() == 0) {
          throw new IOException("Unable to get log information for job: "
              + oldJobID);
        }
        AMInfo amInfo = report.getAMInfos().get(report.getAMInfos().size() - 1);
        return new LogParams(
            amInfo.getContainerId().toString(),
            amInfo.getAppAttemptId().getApplicationId().toString(),
            NodeId.newInstance(amInfo.getNodeManagerHost(),

View on GitHub (pinned to 2add963021)

Solutions

1. Use `yarn logs -applicationId <appId>` (aggregated logs) instead of per-attempt LogParams for finished jobs
2. Verify the attempt actually ran: check the attempt report via JHS REST .../tasks/{taskid}/attempts — attempts with no container cannot yield log paths
3. Retry once JHS has fully loaded the job's history (finish + cache refresh)

Example fix

# before
LogParams lp = cluster.getLogFileParams(oldTaskAttemptID); // IOException

# after
# use the YARN CLI aggregated-log path for finished apps
yarn logs -applicationId application_1400000000000_0001 -logFiles syslog
# or confirm the attempt has a container first via JHS REST attempts listing

TaskAttemptReport ta = jhsRestGetAttempt(jid, tid);
if (ta == null || ta.getContainerId() == null || ta.getNodeManagerHost() == null)
  return Optional.empty(); // no container -> no log params possible

try {
  LogParams lp = cluster.getLogFileParams(attemptId);
} catch (IOException e) {
  if (e.getMessage().startsWith("Unable to get log information for task"))
    return fallbackToYarnLogs(appId); // `yarn logs -applicationId`
  throw e;
}

When it happens

Trigger: Calling getJobLogFileParams/Cluster.getLogFileParams (e.g. via JobClient) for a finished job with a specific oldTaskAttemptID whose attempt never launched a container (was speculatively scheduled then pre-empted, or UNASSIGNED), or whose report from the history server lacks NM location fields.

Common situations: Fetching log parameters for an attempt that failed before container assignment; History server data for the attempt is partial (JHS restarted mid-job, recovery incomplete); Passing an attempt id that belongs to a different run of the job

Related errors

• Unable to get log information for job: {oldJobID}
• User is not set in the application report
• Cannot get log path for a in-progress job
• Unrecognized task type: {}
• Unrecognized State: {}

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/ebd057e43a7a10ea. Report an issue: GitHub.