apache/iceberg · warning · UncheckedIOException

Failed to close the VendedAdlsCredentialProvider

Error message

Failed to close the VendedAdlsCredentialProvider

What it means

VendedAdlsCredentialProvider.close() closes its CloseableGroup (auth manager and client); if closing throws an IOException it is rethrown as this UncheckedIOException. Failures are suppressed per-sub-closeable where possible (setSuppressCloseFailure(true)), so this usually means multiple or unrecoverable close failures.

Source

Thrown at azure/src/main/java/org/apache/iceberg/azure/adlsv2/VendedAdlsCredentialProvider.java:199

  private void checkCredential(Credential credential, String property) {
    Preconditions.checkState(
        credential.config().containsKey(property),
        "Invalid ADLS Credentials: %s not set",
        property);
  }

  @Override
  public void close() {
    CloseableGroup closeableGroup = new CloseableGroup();
    closeableGroup.addCloseable(authSession);
    closeableGroup.addCloseable(authManager);
    closeableGroup.addCloseable(client);
    closeableGroup.setSuppressCloseFailure(true);
    try {
      closeableGroup.close();
    } catch (IOException e) {
      throw new UncheckedIOException("Failed to close the VendedAdlsCredentialProvider", e);
    }
  }
}

View on GitHub (pinned to 86d9c8fc54)

Solutions

1. Inspect the wrapped IOException cause to identify which component failed to close.
2. Ensure token refresh threads are stopped before closing the provider.
3. If thrown during application shutdown, log and continue; it is rarely recoverable, but check for leaked connections/threads in subsequent runs.

try { provider.close(); } catch (UncheckedIOException e) { log.warn("VendedAdlsCredentialProvider close failed: {}", e.getCause().toString()); }

When it happens

Trigger: Calling close() on the VendedAdlsCredentialProvider when the underlying OAuth token fetcher/auth manager or Azure client fails to shut down cleanly.

Common situations: Shutdown during active token refresh; I/O errors during client teardown; resource cleanup in error paths after initialization partially failed.

Understand the failure class

Background: 'Something went wrong' / 'Request failed (500)' / 'HTTP error! status: 404' — what failed HTTP requests actually mean and how to find the real cause — this error's family across 28 libraries.

Related errors

• Location does not exist: %s
• Location already exists: %s
• Failed to create output stream for location:
• Failed to delete path: {}
• Unable to load metrics class: '{}', falling back to null met

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/cee6f95011c783e6. Report an issue: GitHub.