apache/druid · error · ISE

Cannot delete all segment files since Azure Deep Storage sin

Error message

Cannot delete all segment files since Azure Deep Storage since druid.azure.container and druid.azure.prefix are not both set.

What it means

killAll() requires both druid.azure.container and druid.azure.prefix to be configured because it must know where to enumerate and delete all segment blobs. If either is null it throws IllegalStateException instead of attempting a destructive, unscoped operation.

Source

Thrown at extensions-core/azure-extensions/src/main/java/org/apache/druid/storage/azure/AzureDataSegmentKiller.java:136

    Map<String, Object> loadSpec = segment.getLoadSpec();
    final String containerName = MapUtils.getString(loadSpec, "containerName");
    final String blobPath = MapUtils.getString(loadSpec, "blobPath");
    final String dirPath = Paths.get(blobPath).getParent().toString();

    try {
      azureStorage.emptyCloudBlobDirectory(containerName, dirPath);
    }
    catch (BlobStorageException e) {
      throw new SegmentLoadingException(e, "Couldn't kill segment[%s]: [%s]", segment.getId(), e.getMessage());
    }
  }

  @Override
  public void killAll() throws IOException
  {
    if (segmentConfig.getContainer() == null || segmentConfig.getPrefix() == null) {
      throw new ISE(
          "Cannot delete all segment files since Azure Deep Storage since druid.azure.container and druid.azure.prefix are not both set.");
    }
    log.info(
        "Deleting all segment files from Azure storage location [bucket: '%s' prefix: '%s']",
        segmentConfig.getContainer(),
        segmentConfig.getPrefix()
    );
    try {
      AzureUtils.deleteObjectsInPath(
          azureStorage,
          inputDataConfig,
          accountConfig,
          azureCloudBlobIterableFactory,
          segmentConfig.getContainer(),
          segmentConfig.getPrefix(),
          Predicates.alwaysTrue()
      );
    }

View on GitHub (pinned to 9b90983fd2)

Solutions

1. Set druid.azure.container and druid.azure.prefix in the runtime properties before running killAll
2. Verify the AzureDataSegmentConfig picks up the right druid.azure.* keys
3. If killAll is not intended, use per-segment kill instead

Example fix

// before
// druid.storage.type=azure but no druid.azure.* properties
// after
// add to runtime.properties:
// druid.azure.container=my-container
// druid.azure.prefix=druid/segments

if (segmentConfig.getContainer() == null || segmentConfig.getPrefix() == null) {
  throw new IllegalStateException("killAll requires druid.azure.container and druid.azure.prefix");
}

try { killer.killAll(); } catch (ISE e) { log.error("Azure config incomplete for killAll: %s", e.getMessage()); }

When it happens

Trigger: Calling AzureDataSegmentKiller.killAll() when AzureDataSegmentConfig.getContainer() or getPrefix() returns null (these properties not set in the runtime config), typically during a full metadata cleanup task.

Common situations: Running the kill-all/noop cleanup against a cluster where Azure deep storage properties were only partially configured, or using a druid.azure.* config meant for something else.

Understand the failure class

Background: "is required", "must be set", "missing required field": configuration validation errors across open-source libraries — this error's family across 36 libraries.

Related errors

• Set 'account' to the storage account that needs to be config
• Either set 'key' or 'sharedAccessStorageToken' or 'useAzureC
• Failed to parse metric dimensions and types
• Recoverable exception
• NoSuchElementException

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/09aa5f3cb13f7b4b. Report an issue: GitHub.