MetricSign
EN|NLRequest Access
Medium severitydata source

Power BI Refresh Error:
DF-SAPODP-InsufficientResource

What does this error mean?

The SHIR lacks sufficient resources (CPU, memory, or disk) to complete the SAP ODP extraction. Unlike OOM, InsufficientResource is raised when the IR detects approaching resource limits before the process is killed.

Common causes

  • 1The IR host machine is running low on available memory due to other processes or concurrent pipeline runs
  • 2Disk space on the IR host is insufficient to write temporary staging or buffer files
  • 3High CPU usage on the IR host is delaying processing to the point where the connector detects a resource constraint
  • 4Multiple large SAP ODP pipelines are running simultaneously on the same IR host, competing for resources

How to fix it

  1. 1On the Self-Hosted IR host, check current resource usage: open Task Manager and review CPU, memory, and disk usage — identify if any resource is at or near 100%.
  2. 2Reduce pipeline concurrency: if multiple SAP ODP pipelines run simultaneously, schedule them sequentially to reduce peak resource demand.
  3. 3Free up disk space on the IR host — temp files from previous runs, Windows log files, or other applications may be consuming disk.
  4. 4Increase the IR host machine's RAM — 16GB minimum is recommended; 32GB or more for high-volume SAP environments.
  5. 5Lower the rowsPerPage setting in the SAP ODP source to reduce per-run memory consumption.
  6. 6Consider distributing workload across multiple SHIR nodes if the resource constraint is structural.

Frequently asked questions

How is InsufficientResource different from OOM?

OOM means the process was killed for exceeding RAM. InsufficientResource means the connector detected a constraint proactively — the IR may still be running. Both require the same fix: reduce consumption or increase IR capacity.

Can I monitor IR resource usage over time?

Yes — use Windows Performance Monitor (perfmon) on the IR host to log CPU, memory, and disk metrics over time. Run it during pipeline execution to correlate resource peaks with runs — useful for sizing decisions.

Should I add more SHIR nodes to a cluster instead of upgrading hardware?

Adding nodes helps concurrent runs but not a single large extraction bottleneck. Both may be needed: scale up (more RAM per node) for large single extractions; scale out (more nodes) for concurrent pipelines.

Will downstream Power BI datasets be affected?

Yes — the pipeline fails without completing the extraction. Dependent datasets will show stale data.

Official documentation: https://learn.microsoft.com/en-us/azure/data-factory/data-flow-troubleshoot-guide

Other data source errors