apache/hadoop · error · IOException

Can't continue since none or more than one BlockTokenIdentif

Error message

Can't continue since none or more than one BlockTokenIdentifier is found.

What it means

Error "Can't continue since none or more than one BlockTokenIdentifier is found." thrown in apache/hadoop.

Source

Thrown at hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataNode.java:3604

    checkReadAccess(block);
    return data.getReplicaVisibleLength(block);
  }

  private void checkReadAccess(final ExtendedBlock block) throws IOException {
    // Make sure this node has registered for the block pool.
    try {
      getDNRegistrationForBP(block.getBlockPoolId());
    } catch (IOException e) {
      // if it has not registered with the NN, throw an exception back.
      throw new org.apache.hadoop.ipc.RetriableException(
          "Datanode not registered. Try again later.");
    }

    if (isBlockTokenEnabled) {
      Set<TokenIdentifier> tokenIds = UserGroupInformation.getCurrentUser()
          .getTokenIdentifiers();
      if (tokenIds.size() != 1) {
        throw new IOException("Can't continue since none or more than one "
            + "BlockTokenIdentifier is found.");
      }
      for (TokenIdentifier tokenId : tokenIds) {
        BlockTokenIdentifier id = (BlockTokenIdentifier) tokenId;
        LOG.debug("BlockTokenIdentifier: {}", id);
        blockPoolTokenSecretManager.checkAccess(id, null, block,
            BlockTokenIdentifier.AccessMode.READ, null, null);
      }
    }
  }

  /**
   * Transfer a replica to the datanode targets.
   * @param b the block to transfer.
   *          The corresponding replica must be an RBW or a Finalized.
   *          Its GS and numBytes will be set to
   *          the stored GS and the visible length. 
   * @param targets targets to transfer the block to

View on GitHub (pinned to 2add963021)

Solutions

1. Ensure the client sends exactly one block token for the operation; regenerate the access token via the NameNode if the connection carried zero or multiple BlockTokenIdentifiers.
2. Check for custom token serialization code or proxies that duplicate/strip tokens on the data transfer connection.

When it happens

Trigger: During token-based connection setup, the DataNode finds zero or more than one BlockTokenIdentifier among the connection's tokens.

Common situations: Zero or multiple BlockTokenIdentifiers were presented where exactly one is required. Indicates a malformed client request; upgrade mismatched client versions.

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/2f664984d4c8a0ea. Report an issue: GitHub.