March 4, 2024 - Hotfix

Review the known issues, fixed issues, and changed behaviors in this hotfix release of Cloudera Data Warehouse on Public Cloud.

Known issues

This release of the Cloudera Data Warehouse (CDW) service on CDP Public Cloud has the following known issues:

See Data Visualization release notes for known issues in Cloudera Data Visualization 7.1.9.

Carried over from the previous release: Upgrade-related

Upgrading to EKS 1.24 could result in Impala coordinators shutting down.
This issue is not seen on the Impala Virtual Warehouse running Runtime 2023.0.15.0-x or later.
Workaround: Manually start the Impala Virtual Warehouse from the UI or cli. Alternatively, replace your runtime with 1.7.1-b755 (released August 30, 2023) or later.
Certain CDW versions cannot upgrade to EKS 1.22
AWS environments activated in version 1.4.1-b86 (released June, 22, 2022) or earlier are not supported for upgrade to EKS 1.22 due to incompatibility of some components with EKS 1.22. To determine the activation version your pre-existing environment, in the Data Warehouse service, expand Environments. In Environments, search for and locate the environment that you want to view. Click Edit. In Environment Details, you see the CDW version.
DWX-15112 Enterprise Data Warehouse database configuration problems after a Helm-related rollback
If a Helm rollback fails due to an incorrect Enterprise Data Warehouse database configuration, the Virtual Warehouse and Database Catalog roll back to a previous configuration. The incorrect Enterprise Data Warehouse configuration persists, and can affect subsequent edit, upgrade, and rebuild operations on the rolled-back Virtual Warehouse or Database Catalog.

Carried over from the previous release: General

DWX-14923 After JWT authentication, attempting to connect the Impyla client using a user name or password should cause an error
Using a JWT token, you can connect to a Virtual Warehouse as the user who generated the token.
If you connect with the JWT token, and then pass a user name or password from Impyla to the Virtual Warehouse, the connection arguments are silently ignored. Such an action should indicate that it is illegal to specify user or password when using JWT authentication.
DWX-15145 Environment validation popup error after activating an environment
Activating an environment having a public load balancer can cause an environment validation popup error.
To reproduce this problem 1) Create a data lake. 2) Activate an environment having a public load balancer deployment type and subnets in three different availability zones. A environment validation popup can occur even through subnets are in different availability zones. Several different popups can occur, including the following one:
DWX-15144 Virtual Warehouse naming restrictions
You cannot create a Virtual Warehouse having the same name as another Virtual Warehouse even if the like-named Virtual Warehouses are in different environments. You can create a Database Catalog having the same name as another Database Catalog if the Database Catalogs are in different environments.
DWX-13103 Cloudera Data Warehouse environment activation problem
When CDW environments are activated, a race condition can occur between the prometheus pod and istiod pod. The prometheus pod can be set up without an istio-proxy container, causing communication failures to/from prometheus to any other pods in the Kubernetes cluster. Data Warehouse prometheus-related functionalities, such as autoscaling, stop working. Grafana dashboards, which get metrics from prometheus, are not populated.
Workaround: Restart the prometheus pod so that it gets the istio-proxy container.
DWX-5742: Upgrading multiple Hive and Impala Virtual Warehouses or Database Catalogs at the same time fails
Problem: Upgrading multiple Hive and Impala Virtual Warehouses or Database Catalogs at the same time fails.
Workaround: If you need to upgrade or create multiple Hive and Impala Virtual Warehouses or Database Catalogs, perform the upgrade or creation sequentially one at a time.

Carried over from the previous release: AWS

AWS availability zone inventory issue
In this release, you can select a preferred availability zone when you create a Virtual Warehouse; however, AWS might not be able to provide enough compute instances of the type that Cloudera Data Warehouse needs.
Workaround: If you experience this AWS issue, try recreating the Virtual Warehouse and choosing a different availability zone.
DWX-7613: CloudFormation stack creation using AWS CLI broken for CDW Reduced Permissions Mode
Problem: If you use the AWS CLI to create a CloudFormation stack to activate an AWS environment for use in Reduced Permissions Mode, it fails and returns the following error:
An error occurred (ValidationError) when calling the CreateStack
       operation: Parameters: [SdxDDBTableName] must have values
The default value of SdxDDBTableName is not being set. If you create the CloudFormation stack using the AWS Console, there is no problem.
Workaround: If you must use the AWS CLI, edit the CloudFormation stack template file as follows:
SdxDDBTableName:
       Description: DynamoDB table name for the SDX S3 file listings, created through S3Guard
       Type: String
       Default: " "
Then rerun the CloudFormation stack creation command using the AWS CLI.
ENGESC-8271: Helm 2 to Helm 3 migration fails on AWS environments where the overlay network feature is in use and namespaces are stuck in a terminating state
Problem: While using the overlay network feature for AWS environments and after attempting to migrate an AWS environment from Helm 2 to Helm 3, the migration process fails.
Workaround: Run the following kubectl command to determine whether you have any namespaces stuck in a terminating state:
kubectl get ns
Then contact Cloudera Technical Support to report and get help on this issue.
DWX-6970: Tags do not get applied in existing CDW environments
Problem: You may see the following error while trying to apply tags to Virtual Warehouses in an existing CDW environment: An error occurred (UnauthorizedOperation) when calling the CreateTags operation: You are not authorized to perform this operation and Compute node tagging was unsuccessful. This happens because the ec2:CreateTags privilege is missing from your AWS cluster-autoscaler inline policy for the NodeInstanceRole role.
Workaround: Add the ec2:CreateTags privilege to the cluster-autoscaler inline policy as follows:
  1. Log into the AWS IAM console at https://console.aws.amazon.com/iam/.
  2. In the navigation panel, choose Roles.
  3. Search the list of roles for NodeInstanceRole.
  4. Click Permissions.
  5. Select cluster-autoscaler and click Edit policy.
  6. Add the ec2:CreateTags line in the Actions section after the ec2:DescribeLaunchTemplateVersions line as shown:
    "Version": "2012-10-17",
              "Statement": [
              {
              "Action": [
              "autoscaling:DescribeAutoScalingGroups",
              "autoscaling:DescribeAutoScalingInstances",
              "autoscaling:DescribeTags",
              "autoscaling:DescribeLaunchConfigurations",
              "autoscaling:SetDesiredCapacity",
              "autoscaling:TerminateInstanceInAutoScalingGroup",
              "ec2:DescribeLaunchTemplateVersions",
              "ec2:CreateTags"
              ],
  7. Save changes.

Carried over from the previous release: Azure

DWX-17703: Non-HA Impala Virtual Warehouse on a private Azure Kubernetes Service (AKS) setup fails
When 'Refresh' and 'Stop' operations run in parallel, Impala might move into an error state. The Refresh operation might think that Impala is in an error state as the coordinator pod is missing.
Rebuild the Impala Virtual Warehouse or restart it using the CLI.
DWX-17109: ABFS File Browser operations failing intermittently
You may encounter intermittent issues while performing typical operations on files and directories on the ABFS File Broswer, such as moving or renaming files.
None.
DWX-17613: Generic error message is displayed when you click on the directory you don't have access to on a RAZ cluster
You see the following error message when you click on an ABFS directory to which you do not have read/write permission on the ABFS File Browser in Hue: There was a problem with your request. This message is generic and does not provide insight into the actual issue.
None.
DWX-17620: Folder having special characters in its name is not accessible in ABFS
In Cloudera Data Warehouse Public Cloud, from a Virtual Warehouse going to the Azure Blob Filesystem (ABFS), creating a folder, and then performing an action such as Move, causes an error as shown in the following example:
Cannot access: abfs://data-files/user/hrt_qa/~@$&()*!+'=;. 404 Client Error: The specified path does not exist.
DWX-15214 DWX-15176 Hue frontend may become stuck in CrashLoopBackoff on CDW running on Azure
The Hue frontend for Apache Impala (Impala) and Apache Hive (Hive) Virtual Warehouses created in Cloudera Data Warehouse (CDW) can be stuck in a CrashLoopBackoff state on Microsoft Azure (Azure) platform, making it impossible to reach the Virtual Warehouse through Hue. In this case, the following error message is displayed:
kubelet Error: failed to create containerd task: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: exec: “run_httpd.sh”: cannot run executable found relative to current directory: unknown    
This issue affects all CDW releases before 1.6.3 (released May 5, 2023)running on CDP Public Cloud on Azure.
The following users are affected: CDW Azure users using Hue in Impala and Hive Virtual Warehouses in environments earlier than version 1.6.3 or later if they upgrade the node image
Workaround: Upgrade the Virtual Warehouses to 1.6.3 and choose Hue version 2023.0.14.0.

Do not choose Hue versions older than 2023.0.14.0 when creating a Virtual Warehouse in environments of version 1.6.3.

Workloads from earlier CDW versions cannot be deployed on new Azure environments
Because new environments are provisioned automatically to use Azure Kubernetes Service (AKS) 1.22, deprecated APIs used in workloads of earlier versions are not supported in this version 2022.0.9.0-120 (released August 4, 2022). This issue affects you only if you have the CDW_VERSIONED_DEPLOY entitlement.
Workaround: Create new workloads. Workloads in Database Catalogs, Hive, Impala, Hue, Data Visualization, and are incompatible with AKS 1.22.
Incorrect diagnostic bundle location
Problem:The path you see to the diagnostic bundle is wrong when you create a Virtual Warehouse, collect a diagnostic bundle of log files for troubleshooting, and click Edit > Diagnostic Bundle. Your storage account name is missing from the beginning of the path.
Workaround:To find your diagnostic bundle, add your storage account name to the beginning of the path, for example:
my-storage-account-name/log-files/clusters/my-env/my-warehouse-xx-yy/warehouse/debug-artifacts/hive/compute-zz-mvvf/compute-zz-mvvf-0926210111-0927090111-01234.zip
To get your storage account name, in Cloudera Management Console, click Environments, select your environment, and scroll down to Logs Storage and Audits. The first part of the path is your storage account name.
Changed environment credentials not propagated to AKS

Problem: When you change the credentials of a running cloud environment using the Management Console, the changes are not automatically propagated to the corresponding active Cloudera Data Warehouse (CDW) environment. As a result, the Azure Kubernetes Service (AKS) uses old credentials and may not function as expected resulting in inaccessible Hive or Impala Virtual Warehouses.

Workaround: To resolve this issue, you must manually synchronize the changes with the CDW AKS resources. To synchronize the updated credentials, see Update AKS cluster with new service principal credentials in the Azure product documentation.

Carried over from the previous release: Database Catalog

Non-default Database Catalogs created with several earlier CDW versions fails
This issue affects you only if you meet the following conditions:
  • You have a versioned CDW deployment and the multi default DBC entitlement.
  • You are using this release of CDW version 2022.0.9.0-120 (released August 4, 2022).
  • You added Database Catalogs (not the automatically generated default Database Catalog) using either one of these CDW versions:
    • 2022.0.8.0-89 (released June, 22, 2022)
    • 2022-0.7.1-2 (released May 10, 2022)

Do not attempt to upgrade the Database Catalog you created with these earlier releases. The Database Catalog will fail. Your existing Database Catalog created with the earlier release works fine with CDW Runtime. Recreating your existing Database Catalog updates CDW Runtime for 2022.0.9.0-120 (released August 4, 2022).

DWX-7349: In reduced permissions mode, default Database Catalog name does not include the environment name
Problem:

When you activate an AWS environment in reduced permissions mode, the default Database Catalog name does not include the environment name:

This does not cause collisions because each Database Catalog named "default" is associated with a different environment. For more information about reduced permissions mode, see Reduced permissions mode for AWS environments.

Workaround: None available.
DWX-6167: Maximum connections reached when creating multiple Database Catalogs
Problem:After creating 17 Database Catalogs on one AWS environment, Virtual Warehouses failed to start.
Workaround: Limit the number of Database Catalogs created on one environment to 5. This applies to both AWS and Azure environments.

Carried over from the previous release: Hive Virtual Warehouse

Limited Hive image versions
Hive Virtual Warehouses you create in 1.8.1-b248 (released Nov 20, 2023) and later will run Istio 1.19.0. The new Istio version supports only new versions of Hive helm charts. If you have the CDW_VERSIONED_DEPLOY, only new Hive image versions appear in UI when you create a new Hive Virtual Warehouse.
DWX-15064 Hive Virtual Warehouse stops but appears healthy
Due to an istio-proxy problem, the query coordinator can unexpectedly enter a not ready state instead of the expected error-state. Subsequently, the Hive Virtual Warehouse stops when reaching the autosuspend timeout without indicating a problem.
DWX-14452 Parquet table query might fail
Querying a table stored in Parquet from Hive might fail with the following exception message: java.lang.RuntimeException: java.lang.StackOverflowError. This problem can occur when the IN predicate has 243 values or more, and a small stack (-Xss = 256k) is configured for Hive in CDW.
Workaround: Change the value of the Xss in the JVM args to use the default value (1Mb) as follows: Select Edit from the Options menu of your Virtual Warehouse. Click Configurations > Query Executor > and select env.

In the third line shown below, change the value of LLAP_DAEMON_OPTS from -Xss256k to -Xss1M, and then click Apply Changes:

FROM:

-Dorg.wildfly.openssl.path=/usr/lib64 -Dzookeeper.sasl.client=false -Djdk.tls.disabledAlgorithms=SSLv3,GCM -Dhttp.maxConnections=12 -Xms24G -Xmx48G -Xss256k ...

TO:

... -Xss1M ...

Diagnostic bundle download fails
After upgrading to version 1.4.3 (released September 15, 2022), downloading the diagnostic bundle for the Hive Virtual Warehouse results in an error. New environments do not have this problem. The problem is caused by missing permissions to access Kubernetes resources. Version 1.4.3 adds a requirement for the role-based access control (rbac) permissions to download the diagnostic bundle.
Workaround: Add the missing rbac permissions as follows:
1) Create the following file:
apiVersion: rbac.authorization.k8s.io/v1
       kind: ClusterRole
       metadata:
       name: diagnostic-data-generator-role-monitor
       rules:
       - apiGroups:
       - ""
       resources:
       - configmaps
       - events
       - pods
       - persistentvolumeclaims
       - nodes
       verbs:
       - get
       - list
       - apiGroups:
       - apps
       resources:
       - deployments
       - statefulsets
       verbs:
       - get
       - list
       - apiGroups:
       - "edws.cloudera.com"
       resources:
       - computes
       verbs:
       - get
       - list
2) Run the following command to apply the permissions:
kubectl apply -f <filename>
CDPD-40730 Parquet change can cause incompatibility
Parquet files written by the parquet-mr library in this version of CDW, where the schema contains a timestamp with no UTC conversion will not be compatible with older versions of Parquet readers. The effect is that the older versions will still consider these timestamps as they would require UTC conversions and will thus end up with a wrong result. You can encounter this problem only when you write Parquet-based tables using Hive, and tables have the non-default configuration hive.parquet.write.int64.timestamp=true.
DWX-5926: Cloning an existing Hive Virtual Warehouse fails
Problem: If you have an existing Hive Virtual Warehouse that you clone by selecting Clone from the drop-down menu, the cloning process fails. This does not apply to creating a new Hive Virtual Warehouse.
Workaround: Make the following configuration change to resolve this issue:
  1. In the Hive Virtual Warehouse tile, click the edit icon. This launches the Virtual Warehouse details page.
  2. In the details page for the Virtual Warehouse, click the Configurations tab.
  3. Click the Hiveserver2 sub-tab.
  4. Select hive-site from the configuration file drop-down list menu.
  5. Search for the configuration property hive.metastore.sasl.enabled.
  6. Set the hive.metastore.sasl.enabled configuration property to true.
  7. Click Apply in the upper right corner of the page to save the configuration.
  8. Click the Actions menu and select Clone to clone the Hive Virtual Warehouse:
DWX-2690: Older versions of Beeline return SSLPeerUnverifiedException when submitting a query

Problem: When submitting queries to Virtual Warehouses that use Hive, older Beeline clients return an SSLPeerUnverifiedException error:

javax.net.ssl.SSLPeerUnverifiedException: Host name ‘ec2-18-219-32-183.us-east-2.compute.amazonaws.com’ does not
       match the certificate subject provided by the peer (CN=*.env-c25dsw.dwx.cloudera.site) (state=08S01,code=0)

Workaround: Only use Beeline clients from CDP Runtime version 7.0.1.0 or later.

Carried over from the previous release: Hue Query Editor

CDPD-27918: Hue does not automatically pick up RAZ HA configurations
On a CDP Public Cloud environment in which you have configured RAZ in High Availability mode, Hue in CDW does not pick up all the RAZ host URLs automatically. Therefore, if a RAZ instance to which Hue is connected goes down, Hue becomes unavailable.
You must manually add comma-separated RAZ instances in the Hue Advanced Configuration Snippet.
  1. Log in to the CDP Management Console as an Administrator.
  2. Go to Environment > Data Lake and open Cloudera Manager for your environment.
  3. Go to Clusters > Ranger RAZ service > Instances > RAZ server > Processes and note the value of the fs.s3a.ext.raz.rest.host.url property from the core-site.xml file. You need this to specify the value of the api_url property in the Hue configuration.

    For Azure environments, note the value of the fs.azure.ext.raz.rest.host.url property.

    For AWS and GCS environments, note the value of the fs.s3a.ext.raz.rest.host.url property.

  4. Go to CDW > Virtual Warehouse > > Edit > CONFIGURATIONS and select the hue-safety-valve from the Configuration files dropdown menu.
  5. Add the following lines in the hue-safety-valve field:
    [desktop]
    [[raz]]
    is_enabled=true
    api_url=https://[***INSTANCE-1***]:6082/,https://[***INSTANCE-2***]:6082/
  6. Click Apply Changes.
DWX-16899: Error while viewing Impala job status on the mini Job Browser
When you click on the application ID after submitting an Impala query on Hue that is running on the environment level, you may notice the following error: 401 Client Error: Unauthorized for url: http://coordinator.impala-xyz.svc.cluster.local:25000/queries?json=true Must authenticate with Basic authentication. This does not happen when you do the same from Hue deployed at a Virtual Warehouse level.
None.
DWX-16913: “Results have expired” message while running a CTAS query
You may see the following message on the Hue UI when you submit a CREATE TABLE AS SELECT (CTAS) query from a Hue instance that is deployed on the environment level: “Results have expired, rerun the query if needed”. This does not happen when you do the same from Hue deployed at a Virtual Warehouse level.
None.
DWX-16917: Failed to validate proxy privileges error while running queries from Hue
You may see the following error intermittently in Hue’s pod logs while running queries from a Hue instance that is deployed at the environment level: “Failed to validate proxy privileges of <username>”.
None.
DWX-16895: Incorrect status of Hue pods when you edit the Hue instance properties
When you update a configuration of a Hue instance that is deployed at the environment level, such as increasing or decreasing the size of the Hue instance, you see a success message on the CDW UI. After some time, the status of the Hue instance also changes from “Updating” to “Running”. However, when you list the Hue pods using kubectl, you see that not all backend pods are in the running state–a few of them are still in the init state.
None. The pods come up successfully eventually after a sufficient time has passed.
DWX-16863: The upgrade button is present on the CDW UI, but Hue upgrades are not supported
You see the Upgrade button on the Query Editor page in the CDW UI when Hue is deployed at the environment level. However, on CDW version 1.8.1, upgrading the Hue instance that is deployed at the environment level is not supported.
None.
DWX-16893: A user can see all the queries in Job browser
In a Hue instance deployed at the environment level, by design, the Hue instances must not share the saved queries and query history with other Hue instances even for the same user. However, a logged in user is able to view all the queries executed by that user on all the Virtual Warehouses on a particular Database Catalog.
None.
Delay in listing queries in Impala Queries in the Job browser
Listing an Impala query in the Job browser can take an inordinate amount of time.
DWX-14927 Hue fails to list Iceberg snapshots
Hue does not recognize the Iceberg history queries from Hive to list table snapshots. For example, Hue indicates an error at the . before history when you run the following query.
select * from <db_name>.<table_name>.history
DWX-14968 Connection termination error in Impala queries tab after Hue inactivity
To reproduce the problem: 1) In the Impala job browser, navigate to Impala queries. 2) Wait for a few minutes.
After a few minutes of inactivity, the follow error is displayed:
Workaround: Refresh the page, or alternatively start a new session.
DWX-15115 Error displayed after clicking on hyperlink below Hue table browser
In Hue, below the table browser, clicking the hyperlink to a location causes an HTTP 500 error because the file browser is not enabled for environments that are not Ranger authorized (RAZ).
DWX-15090: CSRF error intermittently seen in the Hue Job Browser
You may intermittently see the “403 - CSRF” error on the Hue web interface as well as in the Hue logs after running Hive queries from Hue.
Reload the browser or start a new Hue session.
IMPALA-11447 Selecting certain complex types in Hue crashes Impala
Queries that have structs/arrays in the select list crash Impala if initiated by Hue.
Workaround: Do not select structs/arrays in Hue.
DWX-8460: Unable to delete, move, or rename directories within the S3 bucket from Hue
Problem: You may not be able to rename, move, or delete directories within your S3 bucket from the Hue web interface. This is because of an underlying issue, which will be fixed in a future release.
Workaround: You can move, rename, or delete a directory using the HDFS commands as follows:
  1. SSH into your CDP environment host.
  2. To delete a directory within your S3 bucket, run the following command:
    hdfs dfs -rm -r [***COMPLETE-PATH-TO-S3-BUCKET***]/[***DIRECTORY-NAME***]
  3. To rename a folder, create a new directory and run the following command to move files from the source directory to the target directory:
    hdfs dfs -mkdir [***DIRECTORY-NAME***]
    hdfs dfs -mv [***COMPLETE-PATH-TO-S3-BUCKET***]/[***SOURCE-DIRECTORY***] [***COMPLETE-PATH-TO-S3-BUCKET***]/[***TARGET-DIRECTORY***]
DWX-6674: Hue connection fails on cloned Impala Virtual Warehouses after upgrading
Problem: If you clone an Impala Virtual Warehouse from a recently upgraded Impala Virtual Warehouse, and then try to connect to Hue, the connection fails.
Workaround: Create a new Impala Virtual Warehouse and do not clone from a recently upgraded warehouse. Then the connection to Hue from the new Impala Virtual Warehouse succeeds.
DWX-5650: Hue only makes the first user a superuser for all Virtual Warehouses within a Data Catalog
Problem: Hue marks the user that logs in to Hue from a Virtual Warehouse for the first time as the Hue superuser. But if multiple Virtual Warehouses are connected to a single Data Catalog, then the first user that logs in to any one of the Virtual Warehouses within that Data Catalog is the Hue superuser.

For example, consider that a Data Catalog DC-1 has two Virtual Warehouses VW-1 and VW-2. If a user named John logs in to Hue from VW-1 first, then he becomes the Hue superuser for all the Virtual Warehouses within DC-1. At this time, if Amy logs in to Hue from VW-2, Hue does not make her a superuser within VW-2.

None.

Carried over from the previous release: Iceberg

IMPALA-12742 DELETE/UPDATE Iceberg table partitioned by DATE fails
You can partition Iceberg tables using identities, such as int and date. The file path contains the partition value you can read. When you partition by DATE, subsequently running queries to update or delete data from partitioned data creates an incorrect conversion of DATE. The Catalog cannot parse the data and throws an error.
CDPD-66779: Partitioned Iceberg table not getting loaded with insert select query from Hive
If you create a partitioned table in Iceberg and then try to insert data from another table as shown below, an error occurs.
insert into table partition_transform_4 select t, ts from vectortab10k;
Use the CLUSTER BY clause on the partitioned column to insert data. For example:
insert into table partition_transform_4 select t, ts from t1 cluster by ts;
Branch FAST FORWARD does not work as expected
The Apache Iceberg spec indicates you can use either one or two arguments to fast forward a branch. The following example shows using two arguments:
ALTER TABLE <name> EXECUTE FAST-FORWARD 'x' 'y' 
However, omitting the second branch name, does not work as documented by Apache Iceberg. The named branch is not fast-forwarded to the current branch. An exception occurs at the Iceberg level.
You must use two arguments to the EXECUTE FAST FORWARD feature to forward a branch.
HIVE-28055 Merging Iceberg branches requires a target table alias
Hive supports only one level of qualifier when referencing columns. In other words only one dot is accepted. For example, select table.col from ...; is allowed. select db.table.col is not allowed. Using the merge statement to merge Iceberg branches without a target or source table alias causes an exception:
org.apache.hadoop.hive.ql.parse.SemanticException: ... Invalid table alias or column reference ...
Use an alias, for example t, for the target table.
merge into mydb.target.branch_branch1 t using mydb.source.branch_branch1 s on t.id = s.id when matched then update set value = 'matched';
DWX-17210: Timeout issue querying Iceberg tables from Hive
Workaround: Add the following configuration to hadoop-core-site for the Database Catalog and the Virtual Warehouse.
  • fs.s3.maxConnections=1000
  • fs.s3a.connection.maximum=1000
Restart the Database Catalog and Virtual Warehouse.
DWX-15014 Loading airports table in demo data fails
This issue occurs only when using a non-default Database Catalog. A invalid path error message occurs when using the load command to load the demo data airports table in Iceberg, ORC, or Parquet format.
DWX-14163 Limitations reading Iceberg tables in Avro file format from Impala
The Avro, Impala, and Iceberg specifications describe some limitations related to Avro, and those limitations exist in CDP. In addition to these, the DECIMAL type is not supported in this release.
DEX-7946 Data loss during migration of a Hive table to Iceberg
In this release, by default the table property 'external.table.purge' is set to true, which deletes the table data and metadata if you drop the table during migration from Hive to Iceberg.
Workarounds: Either one of the following workarounds prevents data loss during table migration:
  • Set the table property 'external.table.purge'='FALSE'.
  • Do not drop a table during migration from Hive to Iceberg.
DWX-13733 Timeout issue querying Iceberg tables from Hive
A timeout issue can cause the query to fail.
Workaround: Add the following configuration to hadoop-core-site for the Database Catalog and the Virtual Warehouse.
fs.s3.maxConnections=1000
fs.s3a.connection.maximum=1000
Restart the Database Catalog and Virtual Warehouse.
DWX-13062 Hive-26507 Converting a Hive table having CHAR or VARCHAR columns to Iceberg causes an exception
CHAR and VARCHAR data can be shorter than the length specified by the data type. Remaining characters are padded with spaces. Data is converted to a string in Iceberg. This process can yield incorrect results when you query the converted Iceberg table.
Workaround: Change columns from CHAR or VARCHAR to string types before converting the Hive table to Iceberg.
DWX-13276 Multiple inserts into tables having different formats can cause a deadlock.
Under the following conditions, a deadlock can occur:
  • You run a query to insert data into multiple tables comprised of at least one Iceberg table and at least one non-Iceberg table.
  • The STAT task locking feature is turned on (default = on).
Workarounds: Perform either one of the following workarounds:
  • Run separate queries to insert data into only one table at a time.
  • Turn off STAT task locking as follows:
    set iceberg.hive.request-lock-on-stats-task=false;

Carried over from the previous release: Impala Virtual Warehouse

DWX-17455 ODBC client using JWT authentication cannot connect to Impala Virtual Warehouse
If you are using the Cloudera ODBC Connector and JWT authentication to connect to a CDW Impala Virtual Warehouse where the Impala coordinators are configured for high availability in an active-active mode, the connection results in a "401 Unauthorized" error. The error is not seen on Impala Virtual Warehouses that have an active-passive high availability or where high availability is disabled.
In the ODBC Data Source Name (DSN), make the following changes:
  1. Ensure that SSL is enabled by setting SSL=1.
  2. Remove authentication by setting AuthMech=0.
  3. Remove the JWTString parameter.
  4. Add the parameter http.header.Authorization and set it to the word "Bearer" followed by a space and the JWT string

DSN configuration before the workaround

SSL=1
AuthMech=8
JWTString=full_jwt_text
DSN configuration after the workaround
SSL=1
AuthMech=0
http.header.Authorization=Bearer full_jwt_text
IMPALA-12809 Metadata query problem from Impala
When querying metadata from Impala on a cluster with more than a few nodes, the query might fail with a null pointer exception. Reported problems occurred when running join queries on metadata, but other metadata queries might fail.
Limit the number of nodes that process the query by setting the NUM_NODES query option to 1:
SET NUM_NODES=1
TSB 2023-684: Automatic metadata synchronization across multiple Impala Virtual Warehouses (in CDW Public Cloud) may encounter an exception
The Cloudera Data Warehouse (CDW) Public Cloud 2023.0.14.0 (DWX-1.6.3) version incorporated a feature for performing automatic metadata synchronization across multiple Apache Impala (Impala) Virtual Warehouses. The feature is enabled by default, and relies on the Hive MetaStore events. When a certain sequence of Data Definition Language (DDL) SQL commands are executed as described below, users may encounter a java.lang.NullPointerException (NPE). The exception causes the event processor to stop processing other metadata operations.

If a CREATE TABLE command (not CREATE TABLE AS SELECT) is followed immediately (approximately within 1 second interval) by INVALIDATE METADATA or REFRESH TABLE command on the same table (either on the same Virtual Warehouse or on a different one), there is a possibility that the second command will not find the table in the catalog cache of a peer Virtual Warehouse and generate an NPE.

Knowledge article
For the latest update on this issue see the corresponding Knowledge article: TSB 2023-684: Automatic metadata synchronization across multiple Impala Virtual Warehouses (in CDW Public Cloud) may encounter an exception
DWX-17175: Impala Virtual Warehouse max executor limit issue
The Impala Virtual Warehouse has a limit of max 200 executors. The Executors slider has a max limit of 200 for creating or editing a Virtual Warehouse. If you select a bigger t-shirt size than the default, or set a custom t-shirt size, the default max limit set by the UI could be more than 200 in some cases. Exceeding the max limit causes an error during a future Edit/Upgrade/Rebuild operation.
Workaround: Manually set the max executors limit to a preferred value less than 200, but a multiple of t-shirt size, when creating or editing Impala Virtual Warehouse
DWX-15016 Impala query failure when using Catalog high availability (HA)
Impala queries could fail with InconsistentMetadataFetchException when using HA feature in CDW. This happens when leader election failover for Impala Catalog Service happens due to a transient network error.
Workaround: Disable Catalog HA. In CDW, edit your Virtual Warehouse, and turn off Enable Impala Catalog Server HA.
IMPALA-11045 Impala Virtual Warehouses might produce an error when querying transactional (ACID) table even after you enabled the automatic metadata refresh (version DWX 1.1.2-b2008)
Problem: Impala doesn't open a transaction for select queries, so you might get a FileNotFound error after compaction even though you refreshed the metadata automatically.
Workaround: Run the INVALIDATE METADATA statement on the transactional (ACID) table to refresh the metadata. This fixes the problem until the next compaction occurs. For information about running this statement, see INVALIDATE METADATA statement.
Impala Virtual Warehouses might produce an error when querying transactional (ACID) tables (DWX 1.1.2-b1949 or earlier)
Problem: If you are querying transactional (ACID) tables with an Impala Virtual Warehouse and compaction is run on the compacting Hive Virtual Warehouse, the query might fail. The compacting process deletes files and the Impala Virtual Warehouse might not be aware of the deletion. Then when the Impala Virtual Warehouse attempts to read the deleted file, an error can occur. This situation occurs randomly.
Workaround: Run the INVALIDATE METADATA statement on the transactional (ACID) table to refresh the metadata. This fixes the problem until the next compaction occurs. For information about running this statement, see INVALIDATE METADATA statement.
Do not use the start/stop icons in Impala Virtual Warehouses version 7.2.2.0-106 or earlier
Problem: If you use the stop/start icons in Impala Virtual Warehouses version 7.2.2.0-106 or earlier, it might render the Virtual Warehouse unusable and make it necessary for you to re-create it.
Workaround: Do not use the stop/start icons in these older Virtual Warehouses. Instead, these older versions automatically suspend and resume the Impala executors depending on the absence or presence of queries, making manual start or stop unnecessary.
DWX-6674: Hue connection fails on cloned Impala Virtual Warehouses after upgrading
Problem: If you clone an Impala Virtual Warehouse from a recently upgraded Impala Virtual Warehouse, and then try to connect to Hue, the connection fails.
Workaround: Create a new Impala Virtual Warehouse and do not clone from a recently upgraded warehouse. Then the connection to Hue from the new Impala Virtual Warehouse succeeds.
DWX-5841: Virtual Warehouse endpoints are now restricted to TLS 1.2
Problem: TLS 1.0 and 1.1 are no longer considered secure, so now Virtual Warehouse endpoints must be secured with TLS 1.2 or later, and then the environment that the Virtual Warehouse uses must be reactivated in CDW. This includes both Hive and Impala Virtual Warehouses. To reactivate the environment in the CDW UI:
  1. Deactivate the environment. See Deactivating AWS environments or Deactivating Azure environments.
  2. Activate the environment. See Activating AWS environments or Activating Azure environments
Workaround: If environment reactivation is not possible, you can perform manual steps using the kubectl command line tool to pick up the TLS 1.2 endpoint change. Open a terminal window on a system where the kubectl command line tool is installed, log in, and run the following commands:
kubectl edit svc nginx-service -n <cluster-name>
       
       # Add the following under the metadata.annotations field
       service.beta.kubernetes.io/aws-load-balancer-ssl-negotiation-policy: "ELBSecurityPolicy-TLS-1-2-2017-01"
       
       # Save and quit the editor, and then run the following command to check your changes.
       
       kubectl get svc nginx-service -n <cluster-name> -o yaml
       
       # Make sure that the annotation you added is present.
DWX-5276: Upgrading an older version of an Impala Virtual Warehouse can result in error state
Problem: If you upgrade an older version of an Impala Virtual Warehouse (DWX 1.1.1.1-4) to the latest version, the Virtual Warehouse can get into an Updating or Error state.
Workaround: none
DWX-3914: Collect Diagnostic Bundle option does not work on older environments
The Collect Diagnostic Bundle menu option in Impala Virtual Warehouses does not work for older environments:


Data caching:
This feature is limited to 200 GB per executor, multiplied by the total number of executors.
Sessions with Impala continue to run for 15 minutes after the connection is disconnected.
When a connection to Impala is disconnected, the session continues to run for 15 minutes in case the user or client can reconnect to the same session again by presenting the session_token. After 15 minutes, the client must re-authenticate to Impala to establish a new connection.

UI Issues

DWX-14610 Upgrade icon indicates the Database Catalog is the latest version, but might be incorrect for a few seconds
The cluster fetches data regularly in particular time intervals, or when an action is triggered. Inbetween this time interval, which is very brief, the cluster might not be updated, but will be refreshed in a few seconds.

Technical Service Bulletins

TSB 2023-719: Cloudera Data Warehouse Backup/Restore of CDP Data Visualization incomplete
Cloudera Data Warehouse (CDW) customers using the CDW Automated Backup/Restore feature will encounter an issue with the restoration versions of Cloudera Data Visualization (CDV) older than CDV 7.1.6.2-3 due to a schema change in this release. If the Backup was taken from an older CDW environment that contained a version of CDV older than CDV 7.1.6.2-3, the Restore procedure will succeed. Though once the user opens the CDV Queries tab, the user could encounter the error message: “column jobs_jobschedule.owner_id does not exist…”
Knowledge article
For the latest update on this issue see the corresponding Knowledge article: TSB 2023-719: Cloudera Data Warehouse Backup/Restore of Cloudera Data Visualization incomplete.
TSB 2024-746: Concurrent compactions and modify statements can corrupt Iceberg tables
Apache Hive (Hive) and Apache Impala (Impala) modify statements (DELETE/UPDATE/MERGE) on Apache Iceberg (Iceberg) V2 tables can corrupt the tables if there is a concurrent table compaction from Apache Spark. The issue happens when the compaction and modify statement run in parallel, and when the compaction job commits before the modify statement. In this case the position delete files of the modify statement still point to the old files. This means the following in case of
  • DELETE statements
    • Deleting records pointing to old files have no effect
  • UPDATE / MERGE statements
    • Deleting records pointing to old files have no effect
    • The table will also have the newly added data records
    • Rewritten records will still be active

This issue does not affect Apache NiFi (NiFi) and Apache Flink (Flink) as these components write equality delete files.

Knowledge article

For the latest update on this issue see the corresponding Knowledge article: TSB 2024-746: Concurrent compactions and modify statements can corrupt Iceberg tables.

TSB 2024-752: Dangling delete issue in Spark rewrite_data_files procedure causes incorrect results for Iceberg V2 tables
The Spark Iceberg library includes two procedures - rewrite_data_files and rewrite_position_delete_files. The current implementation of rewrite_data_files has a limitation that the position delete files are not deleted and still tracked by the table metadata, even if they no longer refer to an active data file. This is called the dangling delete problem. To solve this, the rewrite_position_delete_files procedure is implemented in the Spark Iceberg library to remove these old “dangling” position delete files.

Due to the dangling delete limitation, when an Iceberg table with dangling deletes is queried in Impala, Impala tries to optimize select count(*) from iceberg_table query to return the results using stats. This optimization returns incorrect results.

The following conditions must be met for this issue to occur:
  • All delete files in the Iceberg table are “dangling”
    • This would occur immediately after running Spark rewrite_data_files AND
      • Before any further delete operations are performed on the table OR
      • Before Spark rewrite_position_delete_files is run on the table
  • Only stats optimized plain select count(*) from iceberg_table queries are affected. For example, the query should not have:
    • Any WHERE clause
    • Any GROUP BY clause
    • Any HAVING clause

Remove dangling deletes: After rewrite_data_files, position delete records pointing to the rewritten data files are not always marked for removal, and can remain tracked by the live snapshot metadata of the table. This is known as the dangling delete problem.

Knowledge article
For the latest update on this issue see the corresponding Knowledge article: TSB 2024-752: Dangling delete issue in Spark rewrite_data_files procedure causes incorrect results for Iceberg V2 tables.
TSB 2024-758: Truncate command on Iceberg V2 branches cause unintentional data deletion
When working with Apache Hive (Hive) and Apache Iceberg (Iceberg) V2 tables, using the TRUNCATE statement may lead to unintended data deletion. This issue arises when the truncate command is applied to a branch of an Iceberg table. Instead of truncating the branch itself, the command affects the original (main) table, which results in unintended loss of data.
Knowledge article
For the latest update on this issue see the corresponding Knowledge article: TSB 2024-758: Truncate command on Iceberg V2 branches cause unintentional data deletion.

Behavior changes

This release of the Cloudera Data Warehouse (CDW) service on CDP Public Cloud introduces these changes.

This release is a hot fix for a condition related to DWX server code and the Amazon Relational Database (RDS) that causes an inordinate number of events to be saved. Insufficient database connections can occur and Cloudera Data Warehouse can shut down.