Troubleshooting

Note that .HAR files will be required if you raise a support request with Neo4j.

Unable to create account

If you already have an account in Aura, and have used Username-Password or Google Authentication, you are likely to experience this message.

This email is already registered using a different authentication mechanism

Follow the link provided in the error message, or visit https://console.neo4j.io login using Microsoft Login with your Fabric credentials and follow linking instructions.

Unable to link accounts

Users who already have an account in Neo4j Aura may experience problems linking their existing account to their Microsoft account, and may receive the following message

User does not have an email claim. Please contact your Fabric administrator.

Contact your Fabric administrator and ask them to add the email property to your user in the organization’s Microsoft Entra ID. Afterwards, log out, then log in, and attempt to recreate the Graph Dataset.

Queue times for Fabric capacity are too long

Importing data (a.k.a. Transforming OneLake data into a Graph Dataset) requires creating a Fabric Spark environment and then scheduling the import job on your Fabric capacity. The workload tries to perform these steps for you, however the steps heavily depend upon your Fabric capacity’s compute size and readiness. If the importing steps take a long time during environment creation or during capacity queue, try the following:

  • Wait for the capacity queue, rather than cancelling and trying to recreate. Recreating is unlikely to resolve the scheduling/capacity problem.

  • Check the Monitor page (left-hand side panel of Fabric) to see what other jobs are running; they will need to be completed or cancelled before the Graph Dataset can be created.

Failed import jobs

When an import fails, for example during the 'Running Spark job' stage, you have the option to 'Reset DB and retry' or to select the trash can icon to delete the job.

To diagnose the cause of failure, either select Monitor from the left hand Fabric menu, or select the information icon (i) after the stage that failed e.g. 'Running Spark job'. This displays some additional error text and the message To see more detailed information about the error, please visit the monitoring page. CTRL+click the hyperlink to open the Monitor.

Locate the most recent failed import job with your name or use the job id to identify the failed job.

Click on the hyperlinked spark job name e.g. Spark Import Job_a4e3d72e-4bfe-4f62-a1f6-f65aca52dd4e.

Select Logs from the tabs Jobs → Resources → Logs → Data, scroll through the logs until you locate the exception trace to determine the cause of failure. This information will be required if you raise a support request with Neo4j.

Import job fails with TransientException

When there are duplicated values in input data, multiple concurrent to the same nodes can cause lock contention and import job may fail with error like the following:

ExceptionFailure(
  org.neo4j.driver.exceptions.TransientException,
  ForsetiClient[transactionId=<transaction-id>, clientId=<client-id>]
  can't acquire ExclusiveLock{owner=ForsetiClient[transactionId=<transaction-id>, clientId=<client-id>]}
  on NODE(<node-id>) because holders of that lock are waiting for ForsetiClient[...]
)

Review the input data for duplicate values and re-run import job.

Failed Graph Dataset creation

When attempting to add a Graph Dataset you might encounter some error cases which there are known solutions for. Usually this means that the Fabric workspace you are currently using do not have the correct type or permissions.

If you encounter error: There was a problem with the consent, open browser debug console for more details.

This happens because your organization has not granted API permissions to Neo4j Graph Intelligence. Contact your administrator to grant admin permissions to the application API in your organization’s Microsoft Entra ID.

If you encounter the error: Can’t create item in current workspace.

This is because your current Fabric workspace is not a proper Fabric workspace. This usually happens if you are trying to create a Graph Dataset in My Workspace, because My Workspace is a Power BI workspace by default.

You can fix this by assigning a Fabric capacity to the workspace in its workspace settings, either by joining a Fabric Trial or by selecting a capacity that is already available to your organization. The recommended fix, however, is to re-use or create a dedicated Fabric workspace in your organization.

If you get no error but instead the item creation silently fails and nothing happens when you create.

This can be because your Fabric IT Administrator has not enabled "Users can create Fabric items". Ask your IT Administrator to enable the option "Users can create Fabric items" in their "Fabric Admin Portal".

If you encounter the error: Spark environment with this name already exists. Try another name.

Creating a Graph Dataset also creates a Fabric Spark environment with the same name in your workspace. Deleting the Graph Dataset can leave that Spark environment behind, so the name remains in use and the Create button stays disabled.

To continue immediately, enter a different name for the Graph Dataset.

To re-use the original name, delete the leftover Spark environment first:

  1. Close the Create graph dataset dialog and open your Fabric workspace item list.

  2. Locate the item with the name you want to use. Its type is shown as Environment.

  3. Confirm that no other Graph Dataset or notebook depends on it, then delete it.

  4. Create the Graph Dataset again with the original name.