Fixed Issues in Streams Messaging
Review the list of Streams Messaging issues that are resolved in Cloudera DataFlow for Data Hub 7.2.9.
Kafka
There are no fixed Kafka issues in this release.
Streams Messaging Manager
- CDPD-12512: SMM cannot connect to Cloudera Manager metrics service when truststore password is empty
-
SMM throws the following parameter validation error when empty truststore password is used:
Truststore and password must be provided.
Now, an empty truststore password can be used in SMM without an error thrown by the validation in SMM.
- CDPD-22798: Incorrect SMM producer activity
-
A producer in SMM might appear as active minutes after it is no longer active.
- CDPD-23996: No authorization data in logs
-
The authorization data for users does not display in logs.
Schema Registry
- CDPD-21903: Aggregated schema search does not work correctly
- You can now search by schema name and schema description.
Previously, if they tried to search by description only, they would incorrectly get back all schemas in the result set.
- CDPD-23028: Delete model-registry module
- Deleted "model-registry" from the project as it was not upto-date and not used. Removing it allows Cloudera to resolve several other issues.
- CDPD-23183: Introduce Guice and make Atlas an optional feature
- In this release, the framework is replaced by Guice. As a
consequence, you can no longer add custom modules in the
registry.yaml
configuration file. Instead, you can programmatically update by changing the java code.Previously, Schema Registry used a custom-built dependency injection framework.
- CDPD-23996: No authorization data in logs
- The authorization data for users no longer appears in logs.
- OPSAPS-59229: Schema Registry YAML configuration contains plaintext passwords
- Passwords are now assigned as environment variables.
Previously, Schema Registry passwords used to be stored as plaintext in the configuration file.
Streams Replication Manager
- CDPD-22389: The SRM driver role displays as healthy, but replication fails
- The SRM driver role kept running even after it encountered errors that made replication impossible. This resulted in the driver role displaying as healthy in Cloudera Manager even when no replication was done. The driver now automatically stops when such an error is encountered.
- CDPD-22094: The SRM service role displays as healthy, but no metrics are processed
- The SRM service role kept running even after it encountered errors that made metrics processing impossible. This resulted in the service role displaying as healthy in Cloudera Manager even when no metrics were processed. The SRM service role now automatically stops if such an error is encountered.
- CDPD-23683: The replication status reported by the SRM service role for healthy replications is flaky
- The replication status of healthy replications was flaky. The status of the replication could change from active to inactive frequently even if the replication was healthy. The replication status of healthy replications is no longer flaky. The correct (healthy) status is displayed when the replication is healthy.