Known issues and limitations

Learn about the known issues in Flink and SQL Stream Builder, the impact or changes to the functionality, and the workaround in Cloudera Streaming Analytics 1.9.0.

SQL Stream Builder

FLINK-18027: ROW value constructor cannot deal with complex expressions: When querying data from a table or a view with a ROW() function an exception is thrown due to a Calcite parsing issue. For example, the following query will return an error:
CREATE VIEW example AS SELECT col1, ROW(col2) FROM table; SELECT * FROM example;; Add a second SELECT layer to the SQL query as shown in the following example:
CREATE VIEW example AS SELECT col1, ROW(col2) FROM (SELECT col1, col2 FROM table); SELECT * FROM example;
Auto discovery is not supported for Apache Knox: You need to manually configure Knox with SQL Stream Builder to enable Knox authentication.; Complete the configuration based on the CDP Private Cloud Base version you use. For more information, see the Enabling Knox authentication for SSB documentation.
Streaming SQL Console cannot be accessed through Knox when High Availability is enabled: When SQL Stream Builder SSB is deployed in High Availability with Load Balancer, the Streaming SQL Console cannot be accessed directly using Apache Knox.; For accesing Streaming SQL Console, use the secured Load Balancer deployment or authenticate using SPNEGO.
SSB service fails when using Active Directory (AD) Kerberos authentication: If you use AD Kerberos for authentication and the Load Balancer URL is not provided, it can cause the SQL Stream Builder (SSB) service to fail. The issue is caused by the keytab generation. When the keytab is generated by Cloudera Manager it requires the principals from the AD for the Load Balancer host, and without no host specified for the Load Balancer, the SSB service cannot be started by Cloudera Manager. This issue also persists when the Load Balancer role is not deployed or used with SSB.; Fill out the Load Balancer URL parameter in Cloudera Manager regardless of using Load Balancer with SSB. For more information, see the Enabling High Availability for SSB documentation.

Flink

FLINK-18027: ROW value constructor cannot deal with complex expressions: When querying data from a table or a view with a ROW() function an exception is thrown due to a Calcite parsing issue. For example, the following query will return an error:
CREATE VIEW example AS SELECT col1, ROW(col2) FROM table; SELECT * FROM example;; Add a second SELECT layer to the SQL query as shown in the following example:
CREATE VIEW example AS SELECT col1, ROW(col2) FROM (SELECT col1, col2 FROM table); SELECT * FROM example;

In Cloudera Streaming Analytics, the following SQL API features are in preview:

Match recognize
Top-N
Stream-Table join (without rowtime input)

DataStream conversion limitations

Converting between Tables and POJO DataStreams is currently not supported in CSA.
Object arrays are not supported for Tuple conversion.
The java.time class conversions for Tuple DataStreams are only supported by using explicit TypeInformation: LegacyInstantTypeInfo, LocalTimeTypeInfo.getInfoFor(LocalDate/LocalDateTime/LocalTime.class).
Only java.sql.Timestamp is supported for rowtime conversion, java.time.LocalDateTime is not supported.

Kudu catalog limitations

CREATE TABLE
- Primary keys can only be set by the kudu.primary-key-columns property. Using the PRIMARY KEY constraint is not yet possible.
- Range partitioning is not supported.
When getting a table through the catalog, NOT NULL and PRIMARY KEY constraints are ignored. All columns are described as being nullable, and not being primary keys.
Kudu tables cannot be altered through the catalog other than simply renaming them.

Schema Registry catalog limitations

Currently, the Schema Registry catalog / format only supports reading messages with the latest enabled schema for any given Kafka topic at the time when the SQL query was compiled.
No time-column and watermark support for Registry tables.
No CREATE TABLE support. Schemas have to be registered directly in the SchemaRegistry to be accessible through the catalog.
The catalog is read-only. It does not support table deletions or modifications.
By default, it is assumed that Kafka message values contain the schema id as a prefix, because this is the default behaviour for the SchemaRegistry Kafka producer format. To consume messages with schema written in the header, the following property must be set for the Registry client: store.schema.version.id.in.header: true.