Unity Catalog metastore creation fails - Managed Identity could not be found

Zubaer Shaik 0 Reputation points
2026-08-04T01:14:04.4033333+00:00

I am trying to create a Unity Catalog metastore in Azure Databricks.

Error:

Azure Managed Identity Credential with Access Connector Id

/subscriptions/4f453abf-b146-413e-a1a7-9df75573b0ed/resourceGroups/dbricksmeta/providers/Microsoft.Databricks/accessConnectors/spiderconnector

Managed Identity ID:

for Account Id:

eb77f564-0a0e-47e8-9414-cfb3268d5705

could not be found.

I have already verified:

  • Premium workspace
  • Identity Federation enabled
  • Workspace is RUNNING
  • Workspace assigned to metastore
  • Access Connector exists
  • System Assigned Managed Identity exists
  • Managed Identity exists in Microsoft Entra ID
  • Storage Blob Data Contributor assigned
  • Azure CLI confirms the Access Connector identity
  • Databricks CLI confirms the workspace and metastore assignment

Azure CLI output confirms:

principalId: fc79c339-84d6-4ef3-8af5-f782f6b00ae0

tenantId: 643a811a-5688-4b83-af1b-75285edbef5e

provisioningState: Succeeded

Despite this, Databricks reports that the managed identity cannot be found.

Azure Databricks
Azure Databricks

An Apache Spark-based analytics platform optimized for Azure.

0 comments No comments

2 answers

Sort by: Most helpful
  1. Peram Venkateswara Reddy 255 Reputation points Microsoft External Staff Moderator
    2026-08-06T09:15:10.88+00:00

    Hi @Zubaer Shaik , Hope you are doing well.
    Your checklist is thorough, so let me point at the gaps it doesn't cover.

    Your error string contains a clue

    Managed Identity ID:  for Account Id: eb77f564-...
    

    The Managed Identity ID is empty. That's actually correct for your setup — per the docs, that field is only populated "if you created the access connector using a user-assigned managed identity." You're on system-assigned, so leaving it blank is right. This tells us Databricks isn't rejecting a bad value; it can't resolve the connector's system-assigned identity at all. That points at permissions or tenant scope, not configuration.

    Most likely cause: your RBAC on the Access Connector itself

    Your list confirms Storage Blob Data Contributor on the storage account — but that's a different permission from the one needed here. The documentation requires:

    "You must have the Contributor role or higher on the access connector resource in Azure."

    Check the connector's Access control (IAM) → Role assignments and confirm your account (the one signed into the Databricks account console) holds Contributor or Owner on the connector resource, not just on the resource group or storage. If the connector was created by a colleague or by IaC under a service principal, this is very often the missing piece — and it produces exactly a "could not be found" message, because the resolution call returns nothing rather than a permission error.

    First check: tenant alignment

    The error references your Databricks Account Id. Confirm the tenantId returned by

    az databricks access-connector show --name spiderconnector \
      --resource-group dbricksmeta --query identity
    

    matches the Entra tenant your Databricks account (accounts.azuredatabricks.net) is homed in. A connector in a different tenant is invisible to the account even though the CLI shows it as healthy.

    Second: confirm you're an account admin

    Metastore creation happens in the account console, not the workspace, and requires the account admin role. Workspace admin isn't sufficient. Your checklist confirms workspace-level items but not this one.

    On region

    Co-locating the connector, workspace, metastore, and storage is a performance recommendation in the docs, not a validation requirement — so a mismatch is unlikely to be your blocker. Worth noting: only one metastore per region per account, so if one already exists in your target region, creation will conflict.

    If all three check out

    Recreate the Access Connector with a fresh system-assigned identity, re-grant Storage Blob Data Contributor on the storage account, note the new resource ID, and retry. This clears any stale identity metadata. If it still fails with the same message, raise a support case with the timestamp and the (redacted) error — resolution failures against a healthy connector need backend log access.

    References

    Start with the Contributor role on the connector — that's the most common cause of this exact message given everything else you've already verified.

    Kind Regards,
    Microsoft Support Team.

    Was this answer helpful?

    0 comments No comments

  2. Senthil kumar 1,900 Reputation points
    2026-08-04T04:50:56.0033333+00:00

    Hi @Zubaer Shaik

    please confirm are you assigned roles properly from access connector. then only its pass others wise failed in azure databricks.

    Thanks.

    Was this answer helpful?

    0 comments No comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.