How to upgrade SQL Server Big Data Clusters
Applies to: SQL Server 2019 (15.x)
The Microsoft SQL Server 2019 Big Data Clusters add-on will be retired. Support for SQL Server 2019 Big Data Clusters will end on February 28, 2025. All existing users of SQL Server 2019 with Software Assurance will be fully supported on the platform and the software will continue to be maintained through SQL Server cumulative updates until that time. For more information, see the announcement blog post and Big data options on the Microsoft SQL Server platform.
The upgrade path depends on the current version of SQL Server Big Data Cluster. To upgrade from a supported release, including general distribution release (GDR), cumulative update (CU), or quick fix engineering (QFE) update, you can upgrade in place. In-place upgrade from a community technology preview (CTP) or release candidate version of BDC is not supported. You need to remove and recreate the cluster. The following sections describe the steps for each scenario:
The oldest currently supported release of Big Data Clusters is SQL Server 2019 CU8.
Upgrade release notes
Before you proceed, check the upgrade release notes for known issues.
imagePullPolicy was required to be set as
"Always" in the deployment profile control.json file when the cluster was initially deployed. This parameter can't be changed after deployment.
In the case that it is set with a different value, unexpected results may happen during the upgrade process and a cluster redeployment will be required.
Upgrade from supported release
This section explains how to upgrade a SQL Server BDC from a supported release (starting with SQL Server 2019 GDR1) to a newer supported release.
Verify no active Livy sessions.
Make sure no active Livy sessions or batch jobs are running in Azure Data Studio. An easy way to confirm this is either through
curlcommand or a browser to request these URLs:
Back up SQL Server master instance.
Back up HDFS.
azdata bdc hdfs cp --from-path <path> --to-path <path>
azdata bdc hdfs cp --from-path hdfs://user/hive/warehouse/%%D --to-path ./%%D
Update Azure Data CLI (
Follow the instructions for installing Azure Data CLI (
If Azure Data CLI (
azdata) was installed with
pipyou need to manually remove it before installing with the Windows installer or the Linux package manager.
Update the Big Data Cluster.
azdata bdc upgrade -n <clusterName> -t <imageTag> -r <containerRegistry>/<containerRepository>
For example, the following script uses
azdata bdc upgrade -n bdc -t 2019-CU19-ubuntu-20.04 -r mcr.microsoft.com/mssql/bdc
The latest image tags are available at SQL Server 2019 Big Data Clusters release notes.
If you use a private repository to pre-pull the images for deploying or upgrading BDC, ensure that the current build images as well as >the target build images are in the private repository. This enables successful rollback, if necessary. Also, if you changed the >credentials of the private repository since the original deployment, update the corresponding environment variables DOCKER_PASSWORD and >DOCKER_USERNAME. Upgrading using different private repositories for current and target builds is not supported.
Increase the timeout for the upgrade
A timeout can occur if certain components are not upgraded in the allocated time. The following code shows what the failure might look like:
>azdata.EXE bdc upgrade --name <mssql-cluster> Upgrading cluster to version 15.0.4003 NOTE: Cluster upgrade can take a significant amount of time depending on configuration, network speed, and the number of nodes in the cluster. Upgrading Control Plane. Control plane upgrade failed. Failed to upgrade controller.
To increase the timeouts for an upgrade, use --controller-timeout and --component-timeout parameters to specify higher values when you issue the upgrade. This option is only available starting with SQL Server 2019 CU2 release. For example:
azdata bdc upgrade -t 2019-CU19-ubuntu-20.04 --controller-timeout=40 --component-timeout=40 --stability-threshold=3
--controller-timeout designates the number of minutes to wait for the controller or controller db to finish upgrading. --component-timeout designates the amount of time that each subsequent phase of the upgrade has to complete.
To increase the timeouts for an upgrade before the SQL Server 2019 CU19 release, edit the upgrade config map. To edit the upgrade config map:
Run the following command:
kubectl edit configmap controller-upgrade-configmap
Edit the following fields:
controllerUpgradeTimeoutInMinutes Designates the number of minutes to wait for the controller or controller db to finish upgrading. Default is 5. Update to at least 20. totalUpgradeTimeoutInMinutes: Designates the combined amount of time for both the controller and controller db to finish upgrading (controller + controller db upgrade).Default is 10. Update to at least 40. componentUpgradeTimeoutInMinutes: Designates the amount of time that each subsequent phase of the upgrade has to complete. Default is 30. Update to 45.
Save and exit.
Update a BDC deployment from CTP or release candidate
In-place upgrade from a CTP or release candidate build of SQL Server Big Data Clusters is not supported. The following section explains how to manually remove and recreate the cluster.
Backup and delete the old cluster
There is no in place upgrade for big data clusters deployed before SQL Server 2019 GDR1 release. The only way to upgrade to a new release is to manually remove and recreate the cluster. Each release has a unique version of Azure Data CLI (
azdata) that is not compatible with the previous version. Also, if a newer container image is downloaded on cluster deployed with different older version, the latest image might not be compatible with the older images on the cluster. The newer image is pulled if you are using the
latest image tag for in the deployment configuration file for the container settings. By default, each release has a specific image tag corresponding to the SQL Server release version. To upgrade to the latest release, use the following steps:
Before deleting the old cluster, back up the data on the SQL Server master instance and on HDFS. For the SQL Server master instance, you can use SQL Server backup and restore. For HDFS, you can copy out the data with
Delete the old cluster with the
azdata delete clustercommand.
azdata bdc delete --name <old-cluster-name>
Use the version of Azure Data CLI (
azdata) that matches your cluster. Do not delete an older cluster with the newer version of Azure Data CLI (
azdata bdc deletecommand will result in all objects created within the namespace identified with the big data cluster name to be deleted, but not the namespace itself. Namespace can be reused for subsequent deployments as long as it is empty and no other applications were created within.
Uninstall the old version of Azure Data CLI (
pip3 uninstall -r https://azdatacli.blob.core.windows.net/python/azdata/2019-rc1/requirements.txt
Install the latest version of Azure Data CLI (
azdata). The following commands install Azure Data CLI (
azdata) from the latest release:
pip3 install -r https://aka.ms/azdata
pip3 install -r https://aka.ms/azdata --user
For each release, the path to the
n-1version of Azure Data CLI (
azdata) changes. Even if you previously installed Azure Data CLI (
azdata), you must reinstall from the latest path before creating the new cluster.
Verify the azdata version
Before deploying a new big data cluster, verify that you are using the latest version of Azure Data CLI (
azdata) with the
Install the new release
After removing the previous big data cluster and installing the latest Azure Data CLI (
azdata), deploy the new big data cluster by using the current deployment instructions. For more information, see How to deploy SQL Server Big Data Clusters on Kubernetes. Then, restore any required databases or files.
For more information about big data clusters, see What are SQL Server Big Data Clusters.