apache/hadoop · error · YarnRuntimeException

Unrecognized job state: {}

Error message

Unrecognized job state: {}

What it means

TypeConverter.fromYarn(JobState) maps the MRv2 job states (NEW, INITED→PREP; RUNNING; SUCCEEDED; FAILED/ERROR→FAILED; KILLED) onto classic JobStatus run-states and throws YarnRuntimeException for anything unmapped. The MRv2 JobState enum currently has exactly those seven constants, so the throw fires only when a state arrives that this build of the converter does not know — the classic version-skew signature.

Source

Thrown at hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-common/src/main/java/org/apache/hadoop/mapreduce/TypeConverter.java:403

  }


  public static int fromYarn(JobState state) {
    switch (state) {
    case NEW:
    case INITED:
      return org.apache.hadoop.mapred.JobStatus.PREP;
    case RUNNING:
      return org.apache.hadoop.mapred.JobStatus.RUNNING;
    case KILLED:
      return org.apache.hadoop.mapred.JobStatus.KILLED;
    case SUCCEEDED:
      return org.apache.hadoop.mapred.JobStatus.SUCCEEDED;
    case FAILED:
    case ERROR:
      return org.apache.hadoop.mapred.JobStatus.FAILED;
    }
    throw new YarnRuntimeException("Unrecognized job state: " + state);
  }

  public static org.apache.hadoop.mapred.TIPStatus fromYarn(
      TaskState state) {
    switch (state) {
    case NEW:
    case SCHEDULED:
      return org.apache.hadoop.mapred.TIPStatus.PENDING;
    case RUNNING:
      return org.apache.hadoop.mapred.TIPStatus.RUNNING;
    case KILLED:
      return org.apache.hadoop.mapred.TIPStatus.KILLED;
    case SUCCEEDED:
      return org.apache.hadoop.mapred.TIPStatus.COMPLETE;
    case FAILED:
      return org.apache.hadoop.mapred.TIPStatus.FAILED;
    }
    throw new YarnRuntimeException("Unrecognized task state: " + state);

View on GitHub (pinned to 2add963021)

Solutions

  1. Upgrade the client's hadoop-mapreduce-client-common to the cluster's version
  2. During rolling upgrades, keep the client at the highest cluster version
  3. Catch YarnRuntimeException when polling status and report the raw string
Defensive patterns

Strategy: type-guard

Validate before calling

private static final Set<JobState> KNOWN_JOB_STATES = EnumSet.of(
    JobState.NEW, JobState.INITED, JobState.RUNNING, JobState.SUCCEEDED,
    JobState.FAILED, JobState.KILLED, JobState.ERROR);
if (!KNOWN_JOB_STATES.contains(jobState)) { /* raw handling */ }

Type guard

static boolean isMappableJobState(JobState s) {
  return EnumSet.of(JobState.NEW, JobState.INITED, JobState.RUNNING,
      JobState.SUCCEEDED, JobState.FAILED, JobState.KILLED,
      JobState.ERROR).contains(s);
}

Try / catch

try {
  runState = TypeConverter.fromYarn(jobState);
} catch (YarnRuntimeException e) {
  LOG.warn("Unknown job state {} from cluster; treating as RUNNING", jobState);
  runState = JobStatus.RUNNING;
}

Prevention

When it happens

Trigger: A job report from a newer MR AppMaster/JobHistory server carrying a JobState constant not present in this build's switch; mocked JobState values in tests.

Common situations: Down-level mapreduce-client-common talking to an upgraded cluster (rolling upgrade windows are the peak risk); third-party forks adding job states; client tools bundled with an application older than the cluster.

Related errors

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/ec0431aef6b22272. Report an issue: GitHub.