Fixed Issues in Cloudera Manager 7.11.3 Cumulative hotfix 4 for Data Services 1.5.3
Fixed issues in Cloudera Manager 7.11.3 CHF4 (version: 7.11.3.6-50817646) for Data Services 1.5.3 release.
- OPSAPS-69387: Update Spark 3 parcel CSD's repository URL to point to CDP 7.1.9.x cluster in CM
-
Updated the Spark 3 parcel's repository URL to point to https://archive.cloudera.com/p/spark3/3.3.7190.0/parcels/ instead of https://archive.cloudera.com/p/spark3/3.3.7180.0/parcels/.
- OPSAPS-69711: Cloudera Manager - ECS Server host on RHEL 9.1 keeps getting entropy alerts
-
The host level entropy health test turns into
BADstate if the OS version is among RHEL, Cent OS. OEL 9.x. That can causeBADhealth state for the services deployed into these hosts. This issue is fixed now. - OPSAPS-69458: Custom properties atlas.jaas.KafkaClient.option.password appears in a clear text in CDP cluster services.
-
CDP Private Cloud Base 7.1.9 cluster had a configuration property with a clear text password which is a Information security breach. The password is now masked or encrypted in the cluster.
- OPSAPS-69480: Hardcode MR add-opens-as-default config
- Cloudera Manager uses fixed runtime versions when determining clients, instead of using the one connected to the deployed runtime version, which can cause issues. During an upgrade if an app is submitted with a client containing MAPREDUCE-7449 to a runtime that doesn't contain MAPREDUCE-7449's related changes, the application submission fails. To fix this issue MAPREDUCE-7468 changes the default behaviour of the feature to avoid including the placeholder by default. Cloudera Manager has a hardcoded property from the runtime versions where the replacement is correctly done in NM code.
- OPSAPS-69481: Some Kafka connect metrics missing from Cloudera Manager due to conflicting definitions
-
Cloudera Manager now registers
kafka_connect_connector_task_metrics_batch_size_avgandkafka_connect_connector_task_metrics_batch_size_maxmetrics correctly. - OPSAPS-69556: While upgrading from CDP Private Cloud Data Services 1.5.1 to 1.5.2, the public registry with public bits fails with ImagePull Errors, and the docker registry modified to point to docker-private during the upgrade
-
Previously, when upgrading using the Cloudera public registry with public bits, the Docker registry would incorrectly change to point to docker-private.infra.cloudera.com. This issue is now fixed to point to the correct registry.
- OPSAPS-68288: Cloudera Manager waits on "Refreshing Resource manager" during the time when the node-manager is being decommissioned
- The decommission now works as expected.
- OPSAPS-69502: Upgrade failures from CDH6 to 7.1.7 SP3 because ACL is not the expected for znode
- Updated the zk-client.sh to follow the output change of the ZK CLI during upgrade so that the upgrade no longer fails.
