Troubleshooting migration (Watson Knowledge Catalog)

Review the troubleshooting information if you encounter issues during migration.

The migration logs are stored in glossary-service. Use this command to access the logs:
logs wkc-glossary-service-849fdd8cd7-6nq52 -n ${PROJECT_CPD_INSTANCE}

Analyzing imported data and running imported data rules fails

After you import assets, the analysis and data rules execution might fail with an invalid data connection error. The issue might occur because different versions of a connector are used in the source and target systems.

To resolve the issue, reimport metadata on the same imported data connection on the target system in Catalogs > Metadata import.

Trying to view a run history of an imported rule set results in an error

When you’re trying to display the run history for an imported rule set, an error occurs and you’re unable to view the history.

To resolve the issue, reimport the workspace (project) that contains the rule set for which the run history can’t be displayed. For more information, see the sections about importing data quality assets and rule set run history in the Importing the files to the target system topic.

Unable to work with output tables which were added to workspaces as data sets

When you ran data rules in your source environment, you selected the option to save the output tables of the data rules as data sets in workspaces. After data migration, when you try to work with such data sets, an error occurs. In Cloud Pak for Data workspaces are projects and data sets are data assets.

To resolve the issue, complete these steps:
  1. In Cloud Pak for Data, go to Governance > Data quality.
  2. Search for the project that contains this data asset and find the data asset.
  3. Remove the data asset from the project.
  4. Find the data rule which was configured to add the output table as this data asset to the project.
  5. Make sure that the Add the output table as a data asset in this project option is selected in the Output settings section.
  6. Run the data rule.
The output table is again registered as a data asset and you can work with it.

Synchronization of glossary assets fails

The following command that you ran to synchronize assets for display in the Cloud Pak for Data user interface failed.
curl -d -k "https://target-host-name/v3/governance_artifact_types/all/migrate" -H "accept: application/json" -H "Authorization: Bearer your-token"
The status of the migration shows which phase failed. For example:
"Migrate failed"
"Reconcile failed"
"Promote failed"
"SendToGlobalSearch failed"
"Cleanup failed"
The status also contains information about the hours when the synchronization started and when a phase failed.

After you fix the issues that caused the synchronization to fail, you can resume it. For example, if there was not enough space for transaction logs, increase the limit, and then resume the synchronization.

To resume the synchronization from the phase when it failed, run the following command:
curl -k -X GET  "https://target-host-name/v3/governance_artifact_types/all/migrate/resume " -H "accept: application/json" -H "Authorization: Bearer your-token"
After you resume the synchronization, one of the following statuses is displayed:
"Resume phase": "Migrate"
"Resume phase": "Reconcile"
"Resume phase": "Promote"
"Resume phase": "SendToGlobalSearch"
"Resume phase": "Cleanup"
For more information about the commands, see the Synchronizing glossary assets (Watson Knowledge Catalog) topic.

Synchronization of glossary assets and data classes hangs

The synchronization of glossary assets and data classes takes a very long time and doesn’t proceed. The cause for this issue might be that the database XMETA on the metadata repository pod (iis-xmetarepo) is heavily used at the moment. For example, the import of assets might not be complete yet.

To resolve the issue, finish the processes that are running on the XMETA database. The synchronization automatically proceeds without any issues.

Synchronization of glossary assets is incomplete

If glossary assets do not show up in their respective governance artifacts view after you run the migrate command, resynchronize the artifacts by using the following command: 
curl -k -X GET "https://target-host-name/v3/glossary_terms/admin/resync?artifact_type=all" -H "accept: application/json" -H "Authorization: Bearer your-token"

All missing glossary assets (terms, rules, policies) should now be available in their respective views.

Migration fails and HTTP 500 error is thrown

The migration fails, resulting in HTTP 500 with exception: "There is no migration service available".

The XMETA password environment variable is not set (not available) when the glossary pod is starting. In some environments, the XMETA connection is established properly, which means all XMETA environment variables are in place. In other environments, you may experience issues.

The probable cause is a race condition during installation.

Workaround: Restart pod: wkc-glossary-service