SQL requests to Synapse Apach Spark Pool seemingly randomly fails.

Question

We run power bi reports from D365 F&O using a Synapse Apach Spark Pool. We have had no issues with this method till last week when reports started to fail to refresh sometimes. Error messages on the SQL logs are:

When accessing synapse:

Cannot bulk load because the file "---" could not be opened. Operating system error code 12007(failed to retrieve text for this error. Reason: 317).

When running a query:

Cannot process the file "---" because it could not be opened.

The power bi error is below.

Nothing has changed in the way we access or use synapse so why are we experiencing issues now? We have tried to increase to number of cores to medium size but we are still experiencing intermittent failures.

Answer 1

Hello James Li,

Thank you for posting your question in the Microsoft Q&A forum. 

Your intermittent failures with error codes 12007 and file access issues suggest a problem with Synapse's ability to reliably access storage resources. These symptoms typically indicate one of the following underlying issues: 

• Storage Connectivity Problems: The Spark pool is losing connection to your Azure Data Lake Storage (ADLS) account where the files reside. 
• Authentication/Authorization Issues: Temporary failures in the credential chain between Synapse and storage 
• Resource Contention: While you've scaled up cores, there may be throttling at the storage layer 

Few configurations may help to resolve the issue: 

Verify Storage Account Health: 

• Check Azure Service Health for any storage-related incidents 
• Review storage account metrics for throttling or latency spikes 
• Ensure you're not hitting storage account request limits 

Authentication Review: 

• Confirm your linked service credentials haven't expired 
• Check if Managed Identity permissions were modified 
• Test with fresh storage account SAS tokens if used 

Spark Pool Configuration: 

• While you scaled cores, also check: 
• Spark session timeout settings 
• Driver memory allocation 
• Dynamic executor allocation settings 

Network Considerations: 

• Verify no NSG or firewall changes affecting Synapse-storage communication 
• Check if VNet integration is properly configured if used 

Microsoft Support can help investigate the specific error 12007 (typically a DNS resolution or network connectivity issue) if internal troubleshooting doesn't resolve it. 

If the above answer helped, please do not forget to "Accept Answer" as this may help other community members to refer the info if facing a similar issue. Your contribution to the Microsoft Q&A community is highly appreciated. 

Answer 2

Hi James Li

Thanks for the detailed information.

Based on the error message and execution history you have shared, it appears that the issue is most likely related to intermittent DNS resolution failures when the Synapse Spark pool tries to access Delta files in Azure Data Lake (e.g., _delta_log/*.json). Specifically:

• The storage account’s public network access is disabled, which is standard for secure environments.
• However, the private endpoint for the storage account is either misconfigured or not correctly associated with the required Private DNS Zone.
• This results in the Spark pool being unable to resolve the blob endpoint (e.g., *.blob.core.windows.net), leading to intermittent file access failures.

Correct configuration required -

For private endpoint access to function properly, the associated Private DNS Zone must be:

privatelink.blob.core.windows.net

This DNS zone must be -

• Correctly linked to the virtual network used by the Synapse Spark pool.
• Properly resolving the storage account endpoint (e.g., yourstorageaccount.blob.core.windows.net).

Here are some troubleshooting steps that might help you:

Verify Private Endpoint Configuration -

• Navigate to your Storage Account → Networking → Private endpoint connections.
• Check that the private endpoint is linked to the correct DNS zone: privatelink.blob.core.windows.net.

Ensure DNS Zone is Linked to the Spark Pool’s VNet -

• Go to Private DNS Zones in Azure and verify that the zone is linked to the correct Virtual Network.
• Ensure the A-record for your storage account exists and resolves properly.

Optional Test (if needed) - Run this in a Synapse notebook to verify DNS resolution -

%%sh
nslookup <yourstorageaccount>.blob.core.windows.net

Once the private endpoint and DNS configuration are corrected, this should eliminate the DNS-related errors you're seeing (error code 12007).

Please refer this documentation for useful insights - Failed to Restore SQL Managed Instance Database Backup from Azure Blob Storage

I hope this information helps. Please do let us know if you have any further queries.

---

If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.

As your feedback is valuable and can assist others in the community facing similar issues.

Thank you.