What causes "Option value {} for {} is found. It should be an integer" in apache/hadoop?

Use plain integer strings ("6", "3") for both unit-count keys

apache/hadoop · error · IllegalArgumentException

Option value for is found. It should be an integer

Error message

Option value {} for {} is found. It should be an integer

What it means

ECSchema.extractIntOption wraps Integer.parseInt; when a "numDataUnits" or "numParityUnits" value is not a parseable integer (fractional string, letters, empty, embedded whitespace) a NumberFormatException is converted into this IllegalArgumentException naming the offending value and key. The schema never gets built because the counts cannot be interpreted.

Solutions

Use plain integer strings ("6", "3") for both unit-count keys
Trim values and reject/fix fractional formats before constructing the schema
Pre-validate with Integer.parseInt in your own code to fail with a clearer message pointing at the source field

Example fix

// before
opts.put("numDataUnits", "6.0"); // float string
new ECSchema(opts); // throws: It should be an integer

// after
opts.put("numDataUnits", Integer.toString(6));

Defensive patterns

Strategy: validation

Validate before calling

for (String k : new String[]{"numDataUnits", "numParityUnits"}) {
  String v = allOptions.get(k);
  if (v != null) {
    Integer.parseInt(v.trim()); // throws early with your own context
  }
}
new ECSchema(allOptions);

Try / catch

try {
  new ECSchema(allOptions);
} catch (IllegalArgumentException e) {
  Throwable cause = e; // NumberFormatException context is embedded in message
  throw new IllegalArgumentException("Non-integer EC schema option: "
      + allOptions, e);
}

Prevention

Emit integers with Integer.toString when generating policy files; never serialize floats for unit counts
Trim whitespace from values parsed from pretty-printed XML
Reject values matching non-digit patterns at config load time

When it happens

Trigger: new ECSchema(map) with values like "6.0", "3.5", "six", "" or " 6"; JSON/XML numbers serialized as floats or with locale-specific formatting; empty <numDataUnits></numDataUnits> elements.

Common situations: Schema options round-tripped through systems that reformat numbers (floats, thousands separators); hand-edited policy files; whitespace introduced by pretty-printed XML.

Related errors

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/81f9f91393255ea9. Report an issue: GitHub.

Appendix: source

Thrown at hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/erasurecode/ECSchema.java:142

    }

    // After some cleanup
    this.extraOptions = Collections.unmodifiableMap(extraOptions);
  }

  private int extractIntOption(String optionKey, Map<String, String> options) {
    int result = -1;

    try {
      if (options.containsKey(optionKey)) {
        result = Integer.parseInt(options.get(optionKey));
        if (result <= 0) {
          throw new IllegalArgumentException("Bad option value " + result +
              " found for " + optionKey);
        }
      }
    } catch (NumberFormatException e) {
      throw new IllegalArgumentException("Option value " +
          options.get(optionKey) + " for " + optionKey +
          " is found. It should be an integer");
    }

    return result;
  }

  /**
   * Get the codec name
   * @return codec name
   */
  public String getCodecName() {
    return codecName;
  }

  /**
   * Get extra options specific to an erasure code.
   * @return extra options

View on GitHub (pinned to 2add963021)