apache/hadoop · error · YarnRuntimeException

Unrecognized State: {}

Error message

Unrecognized State: {}

What it means

TypeConverter.toYarn(org.apache.hadoop.mapred.TaskStatus.State) maps the classic attempt-state enum onto YARN TaskAttemptState: COMMIT_PENDING, FAILED/FAILED_UNCLEAN, KILLED/KILLED_UNCLEAN, RUNNING, SUCCEEDED, UNASSIGNED(→STARTING). Any other constant hits the default and throws YarnRuntimeException. With the current classic enum all eight constants are covered, so the throw signals an unexpected/new constant (version drift) or a corrupted state value.

Source

Thrown at hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-common/src/main/java/org/apache/hadoop/mapreduce/TypeConverter.java:169

  public static TaskAttemptState toYarn(
      org.apache.hadoop.mapred.TaskStatus.State state) {
    switch (state) {
    case COMMIT_PENDING:
      return TaskAttemptState.COMMIT_PENDING;
    case FAILED:
    case FAILED_UNCLEAN:
      return TaskAttemptState.FAILED;
    case KILLED:
    case KILLED_UNCLEAN:
      return TaskAttemptState.KILLED;
    case RUNNING:
      return TaskAttemptState.RUNNING;
    case SUCCEEDED:
      return TaskAttemptState.SUCCEEDED;
    case UNASSIGNED:
      return TaskAttemptState.STARTING;
    default:
      throw new YarnRuntimeException("Unrecognized State: " + state);
    }
  }

  public static Phase toYarn(org.apache.hadoop.mapred.TaskStatus.Phase phase) {
    switch (phase) {
    case STARTING:
      return Phase.STARTING;
    case MAP:
      return Phase.MAP;
    case SHUFFLE:
      return Phase.SHUFFLE;
    case SORT:
      return Phase.SORT;
    case REDUCE:
      return Phase.REDUCE;
    case CLEANUP:
      return Phase.CLEANUP;
    default:

View on GitHub (pinned to 2add963021)

Solutions

1. Make hadoop-mapreduce-client-core and client-common the same version
2. In tests, use only the real TaskStatus.State constants
3. Catch YarnRuntimeException and surface the raw state string in diagnostics

private static final Set<TaskStatus.State> CONVERTIBLE_STATES = EnumSet.of(
    TaskStatus.State.COMMIT_PENDING, TaskStatus.State.FAILED,
    TaskStatus.State.FAILED_UNCLEAN, TaskStatus.State.KILLED,
    TaskStatus.State.KILLED_UNCLEAN, TaskStatus.State.RUNNING,
    TaskStatus.State.SUCCEEDED, TaskStatus.State.UNASSIGNED);
if (!CONVERTIBLE_STATES.contains(state)) { /* skip or log */ }

static boolean isConvertibleAttemptState(org.apache.hadoop.mapred.TaskStatus.State s) {
  return EnumSet.of(org.apache.hadoop.mapred.TaskStatus.State.COMMIT_PENDING,
      org.apache.hadoop.mapred.TaskStatus.State.FAILED,
      org.apache.hadoop.mapred.TaskStatus.State.FAILED_UNCLEAN,
      org.apache.hadoop.mapred.TaskStatus.State.KILLED,
      org.apache.hadoop.mapred.TaskStatus.State.KILLED_UNCLEAN,
      org.apache.hadoop.mapred.TaskStatus.State.RUNNING,
      org.apache.hadoop.mapred.TaskStatus.State.SUCCEEDED,
      org.apache.hadoop.mapred.TaskStatus.State.UNASSIGNED).contains(s);
}

try {
  attemptState = TypeConverter.toYarn(taskStatus.getState());
} catch (YarnRuntimeException e) {
  LOG.warn("Unmappable attempt state {}", taskStatus.getState());
}

When it happens

Trigger: Converting a TaskStatus whose State is not one of the eight handled constants — practically only possible with a newer mapred enum on the classpath or a hand-crafted TaskStatus in tests.

Common situations: Mixed-version Hadoop jars (newer hadoop-mapreduce-client-core with older client-common); unit tests constructing TaskStatus with custom states; forks that add attempt states.

Related errors

• Unrecognized task type: {}
• Unrecognized Phase: {}
• Unrecognized status: {}
• Unrecognized job state: {}
• Unrecognized task state: {}

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/1ed231e443436120. Report an issue: GitHub.