Skip to main content

Running the Marquez collector on-premise

Ways to run the data.world Collector

There are a few different ways to run the data.world Collector--any of which can be combined with an automation strategy to keep your catalog up to date:

  • Create a configuration file (config.yml) - This option stores all the information needed to catalog your data sources. It is an especially valuable option if you have multiple data sources to catalog as you don't need to run multiple scripts or CLI commands separately.

  • Run the collector through a CLI - Repeat runs of the collector requires you to re-enter the command for each run.

Note

This section walks you through the process of running the collector using CLI.

Preparing and running the command

The easiest way to create your Collector command is to:

  1. Copy the following example command in a text editor.

  2. Set the required parameters in the command. The example command includes the minimal parameters required to run the collector.

  3. Open a terminal window in any Unix environment that uses a Bash shell and paste the command in it and run in.

    Important

    If you are running the collector using Jar files, be sure to edit the command as instructed on this page.

    docker run -it --rm --mount type=bind,source=${HOME}/dwcc,target=/dwcc-output \
  --mount type=bind,source=${HOME}/dwcc,target=/app/log datadotworld/dwcc:2.233 \
  catalog-marquez --agent=8bank-catalog-sources --site=solutions \
  --no-log-upload=false --upload=true --api-token=${DW_AUTH_TOKEN} \
  --name=8bank-catalog-sources-collection --output=/dwcc-output \
  --upload-location=ddw-catalogs --marquez-api-base-url=base_url \
  --marquez-api-key=${DW_MARQUEZ_TOKEN}

    The following table describes the parameters for the command. Detailed information about the Docker portion of the command can be found here.Introduction to Docker

    Table 1.

    Parameter

    Details

    Required?

    dwcc:<CollectorVersion>

    Replace <CollectorVersion> in with the version of the collector you want to use (For example, datadotworld/dwcc:2.233)

    Yes

    -a =<agent>

    --agent=<agent>

    --account=<agent>

    The ID for the data.world account into which you will load this catalog. The ID is the organization name as it appears in your organization. This is used to generate the namespace for any URIs generated.

    Yes

    --marquez-api-base-url=<baseUrl>

    The base URL of the Marquez API.

    Yes

    <apiKey>--marquez-api-key=

    The API key/token for authentication to the Marquez API.

    Yes

    -n=<catalogName>

    --name=<catalogName>

    The name of the catalog - this will be used to generate the ID for the catalog as well as the filename into which the catalog file will be written.

    Yes

    -o=<outputDir>

    --output=<outputDir>

    The output directory into which any catalog files should be written.

    In our example we use the /dwcc -output as it is running in a Docker container and that is what we specified in the script for a Docker mount point.

    You can change this value to anything you would like as long as it matches what you use in the mount point:

    -mount type=bind,source=/tmp,target=/dwcc-output ...-o /dwcc-output

    In this example, the output will be written to the /tmp directory on the local machine, as indicated by the mount point directive. The log file, in addition to any catalog files, will be written to the directory specified in the mount point directive.

    Yes

    --output-name

    Specify the collector output file name to override the default file name. The system automatically adds .dwec.ttl to the end of the provided file name.

    No

    -L

    --no-log-upload

    Do not upload the log of the dwcc run to the organization account's catalogs dataset or to another location specified with --upload-location (ignored if --upload not specified)

    No

    --site=<site>

    The slug for the data.world site into which you will load this catalog this is used to generate the namespace for any URIs generated.

    No

    -H=Host

    --api-host=Host

    The host for the data.world API. NOTE: This parameter is required for single-tenant installations. For example, "api.site.data.world" where "site" is the name of the single-tenant install.

    No

    -t=<apiToken>

    --api-token=<apiToken>

    The data.world API token to use for authentication. The default is to use an environment variable named DW_AUTH_TOKEN.

    No

    -U

    --upload

    Whether to upload the generated catalog to the organization account's catalogs dataset or to another location specified with --upload-location (This requires that the --api-token is specified.)

    No

    --upload-location=<uploadLocation>

    The dataset to which the catalog is to be uploaded, specified as a simple dataset name to upload to that dataset within the organization's account, or [account/dataset] to upload to a dataset in some other account. This parameter is ignored if --upload is not specified.

    No

    --dry-run

    If specified, the collector does not actually harvest any metadata, but just checks the database connection parameters provided by the user and reports success or failure at connecting.

    No



Collector runtime and troubleshooting

The catalog collector may run in several seconds to many minutes depending on the size and complexity of the system being crawled.

  • If the catalog collector runs without issues, you should see no output on the terminal, but a new file that matching *.dwec.ttl should be in the directory you specified for the output.

  • If there was an issue connecting or running the catalog collector, there will be either a stack trace or a *.log file. Both of those can be sent to support to investigate if the errors are not clear.

A list of common issues and problems encountered when running the collectors is available here.

Automating updates to your metadata catalog

Maintaining an up-to-date metadata catalog is crucial and can be achieved by employing Azure Pipelines, CircleCI, or any automation tool of your preference to execute the catalog collector regularly.

There are two primary strategies for setting up the collector run times:

  • Scheduled: You can configure the collector according to the anticipated frequency of metadata changes in your data source and the business need to access updated metadata. It's necessary to account for the completion time of the collector run (which depends on the size of the source) and the time required to load the collector's output into your catalog. This could be for instance daily or weekly. We recommend scheduling the collector run during off-peak times for optimal performance.

  • Event-triggered: If you have set up automations that refresh the data in a source technology, you can set up the collector to execute whenever the upstream jobs are completed successfully. For example, if you're using Airflow, Github actions, dbt, etc., you can configure the collector to automatically run and keep your catalog updated following modifications to your data sources.

In this section: