Databricks on Azure

Destination Azure Databricks #

On this page, you’ll find step-by-step instructions on how to set up your Azure Databricks instance as data target with Arcion. The extracted replicant-cli will be referred to as the $REPLICANT_HOME directory in the following steps.

Prerequisites #

  • A Databricks workspace on Azure
  • Azure container in ADLS Gen2 (Azure Data Lake Storage Gen2)

To create a storage account, see Create a storage account to use with Azure Data Lake Storage Gen2. To create a container, see Create a container.

I. Create a Databricks cluster #

Arcion supports both Databricks all-purpose cluster and SQL Warehouse. The following sections describe how set up each type of cluster.

Set up all-purpose cluster #

If you want to connect to Databricks all-purpose cluster, follow these instructions:

  1. Log in to your Databricks workspace.
  2. From the Databricks console, go to Data Science & Engineering > Compute > Create Compute.
  3. Enter a name for your cluster.
  4. Select the latest Databricks runtime version.
  5. Set up an external stage.
  6. Click Create Cluster.

Get connection details for a cluster #

To establish connection between your Databricks instance and Arcion, you need to provide the connection details for your cluster. You provide these connection details to Replicant using the connection configuration file. To retrieve connection details, follow these steps after you set up a cluster:

  1. Click the Advanced Options toggle.
  2. Click on the JDBC/ODBC tab and take note of the following values:
    • Server Hostname
    • Port
    • JDBC URL

Set up SQL warehouse #

If you want to connect SQL warehouse (SQL Compute), follow these instructions:

  1. Log in to your Databricks workspace.
  2. From the Databricks console, go to SQL > Review SQL warehouses > Create SQL Warehouse.
  3. Enter a name for your SQL warehouse.
  4. Choose cluster size.
  5. Set up an external stage.
  6. Click Create.

Get connection details for a SQL warehouse #

To establish connection between your Databricks instance and Arcion, you need to provide the connection details for your SQL warehouse. You provide these connection details to Replicant using the connection configuration file. To retrieve connection details, follow these steps after you set up a SQL warehouse:

  1. Navigate to the Connection Details tab.

  2. Take note of the following values:

    • Server Hostname
    • Port
    • JDBC URL

II. Create a personal access token for the Databricks cluster #

To create a personal access token, see Generate a personal access token in Databricks documentation. You need this token to configure Arcion Replicant for replication.

III. Configure ADLS container as stage #

To grant Databricks access to ADLS, follow one of the these methods:

The preceding resources use Python to grant access. Instead of Python, you can use Spark configuration properties to access data in Azure storage account.

Spark configuration for cluster #

  1. On the cluster configuration page, click the Advanced Options toggle.
  2. Click the Spark tab.
  3. In the Spark Config textbox, enter your configuration properties.

Spark configuration for SQL warehouse #

  1. Click your username in the top bar of Databricks workspace and select Admin Console from the dropdown.
  2. Click the SQL Warehouse Settings tab.
  3. In the Data Access Configuration textbox, enter your configuration properties.

For example, to access data in Azure storage account using storage account key, enter the following Spark configuration:

fs.azure.account.key.STORAGE_ACCOUNT.dfs.core.windows.net STORAGE_ACCOUNT_KEY

Replace the following:

  • STORAGE_ACCOUNT: your Azure storage account name
  • STORAGE_ACCOUNT_KEY: your storage account key

IV. Obtain the JDBC Driver for Databricks #

Replicant requires the Databricks JDBC Driver as a dependency. To obtain the appropriate driver, follow these steps:

  • Go to the Databricks JDBC Driver download page and download the driver.
  • From the downloaded ZIP, locate and extract the DatabricksJDBC42.jar file.
  • Put the DatabricksJDBC42.jar file inside $REPLICANT_HOME/lib directory.

V. Configure Replicant connection for Databricks #

In this step, you need to provide the Databricks connection details to Arcion. To do so, follow these steps:

  1. You can find a sample connection configuration file databricks.yaml in the $REPLICANT_HOME/conf/conn/ directory.

  2. The connection configuration file has the following two parts:

    • Parameters related to target Databricks server connection.
    • Parameters related to stage configuration.

    Note: All communications with Databricks happen through port 443, the standard port for HTTPS. So all data is encrypted and secure with SSL by default.
    You can store your connection credentials in a secrets management service and tell Replicant to retrieve the credentials. For more information, see Secrets management. Otherwise, you can put your credentials in plain form like the following sample:

    type: DATABRICKS_DELTALAKE

url: "JDBC_URL"
username: USERNAME
password: "PASSWORD"
host: "HOSTNAME"
port: "PORT_NUMBER"
max-connections: 30

    Replace the following:

    Change the value of max-connections as you need. It specifies the maximum number of connections Replicant can open in Databricks.

    For Databricks Unity Catalog, set the connection type to DATABRICKS_LAKEHOUSE. For more information, see Databricks Unity Catalog Support.

    You must use an external stage to hold the data files and load that data on the target database from there. The stage section contains the details Replicant needs to connect to and use a specific stage.

    • type[v21.06.14.1]: The stage type. For Azure Legacy Databricks, set type to AZURE.
    • root-dir: The directory under ADLS container. Replicant uses this directory to stage bulk-load files.
    • conn-url[v21.06.14.1]: The name of the ADLS container.
    • account-name[v21.06.14.1]: The name of the ADLS storage account. account-name corresponds to the same storage account in the Configure ADLS container as stage section.
    • secret-key[v21.06.14.1]: If you want to authenticate ADLS using a storage account key, specify your storage account key here.
    • sas-token: If you use shared access signature (SAS) token to authenticate ADLS, specify the SAS token here.

    The following illustrates two sample stage configurations for Azure Databricks. One sample specifies storage account key and the other sample specifies SAS token for authentication.

    stage:
  type: AZURE
  root-dir: "replicate-stage/databricks-stage"
  conn-url: "replicant-container"
  account-name: "replicant-storageaccount"
  secret-key: "YOUR_STORAGE_ACCOUNT_KEY"

    stage:
  type: AZURE
  root-dir: "replicate-stage/databricks-stage"
  conn-url: "replicant-container"
  account-name: "replicant-storageaccount"
  sas-token: "YOUR_SAS_TOKEN"

VI. Configure mapper file (optional) #

If you want to define data mapping from your source to Azure Databricks, specify the mapping rules in the mapper file. For more information on how to define the mapping rules and run Replicant CLI with the mapper file, see Mapper configuration and Mapper configuration in Databricks.

VII. Set up Applier configuration #

  1. From $REPLICANT_HOME, navigate to the applier configuration file:

    vi conf/dst/databricks.yaml

  2. The configuration file has two parts:

    • Parameters related to snapshot mode.
    • Parameters related to realtime mode.

    For snapshot mode, make the necessary changes as follows:

    snapshot:
  threads: 16 #Maximum number of threads Replicant should use for writing to the target

  #If bulk-load is used, Replicant will use the native bulk-loading capabilities of the target database
  bulk-load:
    enable: true
    type: FILE
    serialize: true|false #Set to true if you want the generated files to be applied in serial/parallel fashion

    There are some additional parameters available that you can use in snapshot mode:

    snapshot:
  enable-optimize-write: true
  enable-auto-compact:  true
  enable-unmanaged-delta-table: false
  unmanaged-delta-table-location:
  init-sk: false
  per-table-config:
    init-sk: false
    shard-key:
    enable-optimize-write: true
    enable-auto-compact: true
    enable-unmanaged-delta-table: false
    unmanaged-delta-table-location:

    These parameters are specific to Databricks as destination. More details about these parameters are as follows:

    • enable-optimize-write: Databricks dynamically optimizes Apache Spark partition sizes based on the actual data, and attempts to write out 128 MB files for each table partition. This is an approximate size and can vary depending on dataset characteristics.

      Default: By default, this parameter is set to true.

    • enable-auto-compact: After an individual write, Databricks checks if files can be compacted further. If so, it runs an OPTIMIZE job to further compact files for partitions that have the most number of small files. The job is run with 128 MB file sizes instead of the 1 GB file size used in the standard OPTIMIZE.

      Default: By default, this parameter is set to true.

    • enable-unmanaged-delta-table: An unmanaged table is a Spark SQL table for which Spark manages only the metadata. The data is stored in the path provided by the user. So when you perform DROP TABLE <example-table>, Spark removes only the metadata and not the data itself. The data is still present in the path you provided.

      Default: By default, this parameter is set to false.

    • unmanaged-delta-table-location: The path where data for the unmanaged table is to be stored. It can be a Databricks DBFS path (for example FileStore/tables), or an S3 path (for example, s3://replicate-stage/unmanaged-table-data) where the S3 bucket is accessible to Databricks.

    • init-sk: Partition-key on the source table is represented as a shard-key by replicant. By default the target table does not include this sharding information. If init-sk is true we add the shard-key/partition key to target table create SQL. Shard-key replication is disabled by default because DML replication with partitioned tables in Databricks is very slow if the partition key has a high distinct count.

      Default: By default, this parameter is set to false.

    • per-table-config: This configuration allows you to specify various properties for target tables on a per table basis.

      • init-sk: Partition-key on the source table is represented as a shard-key by replicant. By default, the target table does not include this sharding information. If init-sk is true we add the shard-key/partition key to target table create SQL. Shard-key replication is disabled by default because DML replication with partitioned tables in\ databricks is very slow if the partition key has a high distinct count.

        Default: By default, this parameter is set to false.

      • shard-key: Shard key to be used for partitioning the target table.

      • enable-optimize-write: Databricks dynamically optimizes Apache Spark partition sizes based on the actual data, and attempts to write out 128 MB files for each table partition. This is an approximate size and can vary depending on dataset characteristics.

        Default: By default, this parameter is set to true.

      • enable-auto-compact: After an individual write, Databricks checks if files can be compacted further. If so, it runs an OPTIMIZE job to further compact files for partitions that have the most number of small files. The job is run with 128 MB file sizes instead of the 1 GB file size used in the standard OPTIMIZE.

        Default: By default, this parameter is set to true.

      • enable-unmanaged-delta-table: An unmanaged table is a Spark SQL table for which Spark manages only the metadata. The data is stored in the path provided by the user. So when you perform DROP TABLE <example-table>, Spark removes only the metadata and not the data itself. The data is still present in the path you provided.

        Default: By default, this parameter is set to false.

      • unmanaged-delta-table-location: The path where data for the unmanaged table is to be stored. It can be a Databricks DBFS path (for example FileStore/tables), or an S3 path (for example, s3://replicate-stage/unmanaged-table-data) where the S3 bucket is accessible to Databricks.

    If you want to operate in realtime mode, you can use the realtime section to specify your configuration. For example:

    realtime:
  threads: 4

  txn-size-rows: 100000
  _traceDBTasks: true
  skip-tables-on-failures : false
  enable-dependency-tracking: true

  reuse-temp-table: true

    Additional parameters #

    reuse-temp-table

    true or false.

    Enables reusing temporary tables instead of creating and dropping each time the Applier ingests CDC operations.

    The Applier creates temporary table using target table schema to ingest CDC operations to target table. Creating and dropping temporary tables on each CDC replay may slow down CDC ingestion. Enabling this parameter allows you to make CDC ingestion faster by telling the Applier to reuse the temporary tables.

    Default: true.

    parallel-transaction [v23.06.30.2]

    This configuration enables parallel batch processing. Parallel batch processing processes large transactions by splitting them into multiple batches and processing the batches concurrently. This speeds up real-time replication for large transactions and improves overall performance. This feature is available for both Legacy Databricks and Unity Catalog.

    The following configuration options are available:

    Option Value Details
    enable Boolean. {true|false}. Enables parallel batch processing.

    Default: true when txn-size-rows is greater than or equal to 2_000_000.
    txn-rows-threshold Integer Sets the threshold limit for a transaction to qualify for splitting. If transaction size hits this threshold limit, the Applier thread splits the transaction into multiple batches for parallel processing.

    Default: 1_000_000.
    max-chunks-per-table Integer Sets the number of batches to split a transaction into for parallel processing.

    Default: 5.
    threads Integer Sets the number of threads responsible for parallel batch processing. For example, if the maximum number of chunks for each table is 5 and threads is set to 10, the Applier processes only two transactions concurrently at a time. If new transactions come simultaneously, the main Applier threads process them serially in one batch.

    Default: The number of available processors to JVM.

    Enable Type-2 CDC #

    From version 22.07.19.3 onwards, Arcion supports Type-2 CDC for Databricks as the target. For more information, see Type-2 CDC and cdc-metadata-type.

For a detailed explanation of configuration parameters in the Applier file, read Applier Reference.

Databricks Unity Catalog support #

From version 22.08.31.3 onwards, Arcion supports Databricks Unity Catalog.

As of now, note the following about the state of Arcion’s Unity Catalog support:

  • Legacy Databricks only supports two-level namespace:

    • Schemas
    • Tables

    With introduction of Unity Catalog, Databricks now exposes a three-level namespace that organizes data.

    • Catalogs
    • Schemas
    • Tables

    Arcion adds support for Unity Catalog by introducing a new child storage type (DATABRICKS_LAKEHOUSE child of DATABRICKS_DELTALAKE).

  • If you’re using Unity Catalog, notice the following when configuring your Target Databricks with Arcion:

    • Set the connection type to DATABRICKS_LAKEHOUSE in the connection configuration file.
    • To avoid manual steps to configure staging, Databricks has introduced personal staging. To read the staging URL, we’ve added a new configuration parameter UNITY_CATALOG_PERSONAL_STAGE. The complete stage configuration is as follows: 
      stage:
  type: UNITY_CATALOG_PERSONAL_STAGE
  staging-url: STAGING_URL
  file-format: DATA_FILE_FORMAT
      Replace the following:

      • STAGING_URL: the temporary staging URL—for example, stage://tmp/userName/rootDir.

      • DATA_FILE_FORMAT: the type of data file format. Supported formats are PARQUET and CSV.

        Default: PARQUET.

  • We use SparkJDBC42 driver for Legacy Databricks (DATABRICKS_DELTALAKE) and DatabricksJDBC42 for Unity catalog (DATABRICKS_LAKEHOUSE). For instructions on how to obtain these drivers, see Obtain the JDBC Driver for Databricks.

  • Replicant supports Unity Catalog on AWS and Azure platforms.

  • To configure Mapper file in Unity Catalog, see Mapping in Unity Catalog.