apache/hadoop · error · UnsupportedOperationException

Not supported

Error message

Not supported

What it means

LocalJobRunner.getLogFileParams(JobID, TaskAttemptID) throws UnsupportedOperationException('Not supported') because the local engine keeps no per-attempt aggregated-log directories or LogParams records — that machinery (MR logs to NM/JobHistory via the MR AM) only exists in cluster mode. The same class also returns empty TaskReport[] and delegates most admin operations to no-ops.

Source

Thrown at hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-common/src/main/java/org/apache/hadoop/mapred/LocalJobRunner.java:1021

  }

  @Override
  public Token<DelegationTokenIdentifier> 
     getDelegationToken(Text renewer) throws IOException, InterruptedException {
    return null;
  }

  @Override
  public long renewDelegationToken(Token<DelegationTokenIdentifier> token
                                      ) throws IOException,InterruptedException{
    return 0;
  }

  @Override
  public LogParams getLogFileParams(org.apache.hadoop.mapreduce.JobID jobID,
      org.apache.hadoop.mapreduce.TaskAttemptID taskAttemptID)
      throws IOException, InterruptedException {
    throw new UnsupportedOperationException("Not supported");
  }
  
  static void setupChildMapredLocalDirs(Task t, JobConf conf) {
    String[] localDirs = conf.getTrimmedStrings(MRConfig.LOCAL_DIR);
    String jobId = t.getJobID().toString();
    String taskId = t.getTaskID().toString();
    boolean isCleanup = t.isTaskCleanupTask();
    String user = t.getUser();
    StringBuilder childMapredLocalDir =
        new StringBuilder(localDirs[0] + Path.SEPARATOR
            + getLocalTaskDir(user, jobId, taskId, isCleanup));
    for (int i = 1; i < localDirs.length; i++) {
      childMapredLocalDir.append("," + localDirs[i] + Path.SEPARATOR
          + getLocalTaskDir(user, jobId, taskId, isCleanup));
    }
    LOG.debug(MRConfig.LOCAL_DIR + " for child : " + childMapredLocalDir);
    conf.set(MRConfig.LOCAL_DIR, childMapredLocalDir.toString());
  }

View on GitHub (pinned to 2add963021)

Solutions

  1. Read the task's stdout/stderr directly from the local task work dir (see the task's system dir / local dirs)
  2. Run the job on YARN to use the JobHistory/NM log service
  3. Skip getLogFileParams when mapreduce.framework.name is local

Example fix

// before
LogParams lp = client.getLogFileParams(jobId, attemptId); // throws locally

// after
if (!"local".equals(conf.get("mapreduce.framework.name", "local"))) {
  LogParams lp = client.getLogFileParams(jobId, attemptId);
} else {
  // read stdout/stderr from the local task dir instead
}
Defensive patterns

Strategy: validation

Validate before calling

if (!"local".equals(conf.get("mapreduce.framework.name", "local"))) {
  LogParams lp = client.getLogFileParams(jobId, attemptId);
} else {
  // locally: read stdout/stderr files from the task's local work dir
}

Try / catch

try {
  return client.getLogFileParams(jobId, attemptId);
} catch (UnsupportedOperationException e) {
  return null; // local mode has no log service; caller falls back to local files
}

Prevention

When it happens

Trigger: Calling JobClient.getLogFileParams(jobId, taskAttemptId), e.g. via 'hadoop job -logs <jobid> <attemptid>', against a locally executed job.

Common situations: Log-fetching utilities or log-collection agents pointed at a local-mode job; debugging a local test with cluster-oriented log commands; tools that unconditionally query log params after job completion.

Related errors

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/e379b025fdac9913. Report an issue: GitHub.