Known Issues in Iceberg REST Catalog

Known issues and technical limitations for Cloudera Iceberg REST Catalog are addressed in Cloudera Runtime 7.3.2, its service packs, and cumulative hotfixes.

Cloudera Runtime 7.3.2.10000 SP1

CDPD-121196: Snowflake fails to retrieve Azure vended credentials
7.3.2.10000 SP1
When Data Sharing is configured on Azure and an external Snowflake engine uses ACCESS_DELEGATION_MODE = VENDED_CREDENTIALS, Snowflake fails to parse the ADLS Gen2 vended credentials and reports SQL Execution Error: Failed to retrieve credentials from the Catalog. The Cloudera Cloudera Iceberg REST Catalog vends the Azure SAS token in the standard storage-credentials array as defined by the Apache Iceberg OpenAPI spec, but Snowflake currently expects it in the top-level config map for Azure paths. As a result, Snowflake data sharing with vended credentials works on AWS but fails on Azure.
Configure Snowflake to use external volume credentials instead. Set the catalog integration's ACCESS_DELEGATION_MODE to EXTERNAL_VOLUME_CREDENTIALS and provide a working EXTERNAL_VOLUME when creating the Iceberg table in Snowflake. This bypasses the credential-vending path entirely while still using the REST catalog for metadata. For more information, see CREATE CATALOG INTEGRATION (Apache Iceberg REST) and Use catalog-vended credentials for Iceberg tables in the Snowflake documentation.
CDPD-119352: REST Catalog port 8090 conflicts with YARN ResourceManager on on-premises clusters
7.3.2.10000 SP1
On on-premises clusters with TLS enabled for the Iceberg REST Catalog, the default REST Catalog port 8090 can conflict with the YARN ResourceManager HTTPS web application port. REST Catalog namespace API calls can return HTTP 307 Temporary Redirect or fail to reach the catalog service.
Change the YARN ResourceManager web application HTTPS port (for example, to 8092) so it does not bind to port 8090, then retry REST Catalog API calls through Knox.
CDPD-119319: External client table reads fail when the client does not request vended credentials
7.3.2.10000 SP1
When an external engine (for example, EMR or standalone Spark) reads a shared table through the Iceberg REST Catalog, the client must explicitly request vended credentials by sending the X-Iceberg-Access-Delegation: vended-credentials HTTP header. If the client does not send this header, the REST Catalog assumes the client authenticates to object storage with its own ambient or local credentials. In that case the REST Catalog load-table and /credentials API calls still succeed, but the client falls back to its environment credentials (for example, the AWS EC2 instance role of the Spark host) when reading table data from S3. Because that identity is not authorized on the shared bucket path, the object-storage read fails with HTTP 403 Forbidden. This was originally observed on AWS mixed-mode QI 95 deployments, but the behavior is not specific to mixed-mode or TLS configuration. QI 95 refers to the Cloudera Encryption on wire everywhere (Encryption of Data-in-Motion) quality initiative. A mixed-mode QI 95 deployment is a cluster in which the Data Lake and Data Hub are configured with different TLS versions.
Ensure the external client requests vended credentials by sending the X-Iceberg-Access-Delegation: vended-credentials header, and use an Iceberg client version that sets this header (for example, Iceberg 1.10.2). This behavior conforms to the Iceberg REST Catalog OpenAPI specification.

Known issues identified in Cloudera Runtime 7.3.2.100 CHF 1

There are no new known issues in this release.

Known issues identified in Cloudera Runtime 7.3.2

CDPD-79748: Client ID and secret can access namespaces and DB metadata without creating a data share
7.3.2 and higher service packs and cumulative hotfixes
A client ID and secret can be used to call namespace/database APIs and retrieve database metadata even when no Data Share is created. The behavior is tied to Ranger-side policy configuration and can expose namespace or database details beyond intended share boundaries.
Review Ranger default policies and remove broad public access for external users. Configure deny and exclude policy rules so only explicitly allowed share members can access shared data.
CDPD-80346: Client region provisioning for cloud credentials
7.3.2 and higher service packs and cumulative hotfixes
For S3-backed Iceberg access, the IDBroker credential response does not include client region information. As a result, Data Sharing relies on a configured fallback client region, which can lead to issues for cross-region access and multi-region resilience scenarios.
Configure client_region in hive-site.xml for the target environment.
CDPD-90155: DataShareAdmin has access to all tables in the Data Lake
7.3.2 and higher service packs and cumulative hotfixes
A user assigned the DataShareAdmin role can access all tables in the data lake without explicit grants from a data owner. This behavior is tied to broad Ranger and Knox admin privileges inherited by the DataShareAdmin role and can allow unintended data sharing, which impacts separation-of-duties and least-privilege governance expectations.
None
CDPD-91957: Accessing a Hive table from an external client returns an error on non-us-west-2 environments
7.3.2 and higher service packs and cumulative hotfixes
When an environment is located in a region other than us-west-2, accessing a Hive table from an external client (such as AWS Elastic MapReduce or Standalone Spark) using the REST catalog might return a 301 redirect error. This occurs because the client.region value defaults to us-west-2 when not explicitly configured, and attempting to access an Amazon S3 file with a client configured for a different region results in an error.
Manually configure the client.region value in the Hive Metastore configuration for the target environment after enabling the Cloudera Iceberg REST Catalog.

Known issues identified before Cloudera Runtime 7.3.2

Known issues identified before Cloudera Runtime 7.3.2 include only unresolved issues from previous releases that continue to affect the Cloudera Runtime 7.3.2 base release.

CDPD-81718: Check Metering V2 Service health before starting Rest Catalog Jetty Server
7.3.1.400 and higher service packs and cumulative hotfixes
Checking the Metering V2 Service health status before starting the REST Catalog service has been temporarily removed. This check was causing failures in environments where the Metering Service is not available, such as certain Cloudera Manager-based environments.
None
CDPD-83244: Snowflake Catalog Integration Secret Rotation
7.3.1.400 and higher service packs and cumulative hotfixes
When creating a Snowflake catalog integration, Snowflake does not allow changing the CLIENT ID for the integration, only the secret. However, when using Knox to generate these credentials, both a new CLIENT ID and a new secret are generated. Replacing the catalog integration is not possible if it has active tables.
Change the catalog integration for each table individually or create a new table. Alternatively, you can use a workaround by altering the catalog integration with only the new secret for the old CLIENT ID, provided you have the correct Ranger policy for the new CLIENT ID.
CDPD-83243: Snowflake Catalog Integration with Vended Credentials is not supported
7.3.1.400 and higher service packs and cumulative hotfixes
Snowflake catalog integration with vended credentials is not supported. Attempting to use the /credentials API endpoint results in a No route to host or BadRequestException error, and table creation fails.
None
CDPD-82198: Metering events are sent for deleted or unauthorized tables
7.3.2
7.3.1.400 and higher service packs and cumulative hotfixes
Metering events are generated for deleted or unauthorized tables. If API calls are made for tables that do not exist or fail on the server side (for example, the user is not authorized on the table, or an error occurs while processing the request on the backend in the Hive Metastore, Apache Ranger, or Apache Knox), these calls are still metered and billed. This is expected behavior.
None