Power BI Refresh Error:
UserErrorSHIRHighMemoryLimit

What does this error mean?

The SHIR node ran out of memory during a large data copy or complex data flow execution. ADF terminated the activity to protect the host machine.

Common causes

• 1A single copy activity is loading too much data into memory without partitioning
• 2Multiple pipeline activities are running concurrently on the same SHIR node, exhausting available RAM
• 3The SHIR host machine is undersized for the data volumes being processed
• 4A data flow is running in interpreted mode on the SHIR host instead of being offloaded to Spark
• 5Memory leak in an older version of the SHIR software

How to fix it

1. 1Add partitioning to the copy activity — split large tables by a partition column or date range.
2. 2Limit the number of concurrent activities on the SHIR node using the 'concurrent jobs' setting in IR configuration.
3. 3Upgrade the SHIR host machine to a larger VM with more RAM (minimum 16 GB recommended for large data volumes).
4. 4Update the SHIR software to the latest version to benefit from memory management improvements.
5. 5For data flows, consider switching to an Azure IR with a Spark cluster to offload memory-intensive transformations from the SHIR host.
6. 6Schedule resource-intensive pipelines during off-peak hours to avoid memory contention with other activities.

Frequently asked questions

Other infrastructure errors