Vnet Gateway issues on Power bi Conn

[ChsAIkrishna]

Team,

We are getting frequent vnet gateway failures on power bi Dataset using DAX(simple DAX not complex) and upon the rerun it is working, is any permanent fix for this ?

Error :

{"error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","parameters":{},"details":[{"code":"DM_ErrorDetailNameCode_UnderlyingErrorCode","detail":{"type":1,"value":"-2147467259"}},{"code":"DM_ErrorDetailNameCode_UnderlyingErrorMessage","detail":{"type":1,"value":"Evaluation ran out of memory and can't continue."}},{"code":"DM_ErrorDetailNameCode_UnderlyingHResult","detail":{"type":1,"value":"-2147467259"}}],"exceptionCulprit":1}}}
----------------------------
Query execution failed for dataset 'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX'.
----------------------------
An error has occurred during report processing.
----------------------------
An error occurred during local report processing.

Error 2 

ERROR [HY000] [Microsoft][DriverSupport] (60000) Arrow Based Set: An error occurred while attempting to decompress LZ4 frames. The LZ4 API function LZ4F_decompress returned an error. Error Detail: ERROR_allocation_failed"}},

Error 3

{"error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","parameters":{},"details":[{"code":"DM_ErrorDetailNameCode_UnderlyingErrorCode","detail":{"type":1,"value":"-2147467259"}},{"code":"DM_ErrorDetailNameCode_UnderlyingErrorMessage","detail":{"type":1,"value":"ERROR [HY000] [Microsoft][DriverSupport] (60000) Arrow Based Set: An error occurred while attempting to decompress LZ4 frames. The LZ4 API function LZ4F_decompress returned an error. Error Detail: ERROR_allocation_failed"}},{"code":"DM_ErrorDetailNameCode_UnderlyingHResult","detail":{"type":1,"value":"-2147467259"}},{"code":"Microsoft.Data.Mashup.ErrorCode","detail":{"type":1,"value":null}},{"code":"Microsoft.Data.Mashup.ValueError.Reason","detail":{"type":1,"value":"DataSource.Error"}}],"exceptionCulprit":1}}}
{"error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","parameters":{},"details":[{"code":"DM_ErrorDetailNameCode_UnderlyingErrorCode","detail":{"type":1,"value":"-2147467259"}},{"code":"DM_ErrorDetailNameCode_UnderlyingErrorMessage","detail":{"type":1,"value":"ERROR [HY000] [Microsoft][DriverSupport] (60000) Arrow Based Set: An error occurred while attempting to decompress LZ4 frames. The LZ4 API function LZ4F_decompress returned an error. Error Detail: ERROR_allocation_failed"}},{"code":"DM_ErrorDetailNameCode_UnderlyingHResult","detail":{"type":1,"value":"-2147467259"}},{"code":"Microsoft.Data.Mashup.ErrorCode","detail":{"type":1,"value":null}},{"code":"Microsoft.Data.Mashup.ValueError.Reason","detail":{"type":1,"value":"DataSource.Error"}}],"exceptionCulprit":1}}}
----------------------------
Query execution failed for dataset 'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX'.
----------------------------
An error has occurred during report processing.
----------------------------
An error occurred during local report processing.

[F_Goudarzi]

Similar issue. Any solution?

[mark_ott]

Frequent VNet gateway errors in Power BI related to “DM_GWPipeline_Gateway_MashupDataAccessError” and memory allocation issues often stem from resource limits, configuration problems, or inefficient modeling—even with simple DAX. No single “permanent fix” exists, but several specific corrective actions can significantly reduce error frequency and improve reliability.

Key Causes

• Power BI query memory limit: Every query runs within specific capacity memory boundaries, which depend on your Power BI service tier (Pro, Premium, Fabric, etc.). Even simple queries can hit limits if the model or report is poorly optimized or if there’s resource contention at the time of execution.​

• VNet gateway configuration and connectivity: Outages, misconfigurations, or periodic resource constraints (e.g., spikes in network or gateway usage) can cause intermittent errors. Sometimes the underlying data source is inaccessible from the VNet or runs out of memory or compute.​

• Data size, inefficient queries, and modelling: Models with lots of relationships, large tables, or columns with high cardinality increase load on both gateway and memory—even if DAX is simple.​

• LZ4 decompression failures: These point to memory allocation failures during data transfer or processing, often related to gateway machine resources or configuration.​

Steps to Fix & Mitigate

• Optimize Data Model in Power BI:

  • Reduce dataset size by removing unused columns/tables and using efficient data types.

  • Minimize cardinality in relationships and avoid excessive calculated columns/measures.​

  • Filter imported data so only what’s needed is loaded.

• Review Gateway Machine Resources:

  • Ensure the gateway VM or server has adequate RAM and CPU. Consider increasing capacity if running multiple reports/refreshes.

  • Restart or scale up the gateway host, especially if memory usage accumulates over time.​

• Check VNet Gateway and Data Source Connectivity:

  • Confirm the data source is reliably accessible from the VNet, not experiencing intermittent outages or latency spikes.​

  • Ensure any credential or permission updates on the source are reflected in the gateway configuration.

• Modify Capacity Settings:

  • For Premium/Fabric, review and possibly increase capacity or adjust “Query Memory Limit %” if lowered from the default.​

  • For service refresh issues, spread loads (refresh schedules) to avoid peak times.

• Update Power BI and Gateway Software:

  • Ensure both Power BI Desktop, gateway software, and drivers are the latest production versions.

• Troubleshoot LZ4 ERROR_allocation_failed:

  • Confirm the gateway server has ample free memory available for decompression tasks.

  • Reboot the gateway server if allocation failures persist, and monitor resource usage during refreshes.​

Additional Notes

• The error often disappears upon rerunning because temporary conditions (resource spikes, network hiccups) clear. However, frequent recurrence suggests underlying capacity or configuration needs attention.​

• These memory and allocation errors are not unique to complex DAX; inefficient models or stressed infrastructure can cause them with any DAX formula.

If issues persist after these steps, consider further capacity upgrades or consult Microsoft support with full environmentent details.