What causes UserErrorParquetJavaInvocationException?

The _JAVA_OPTIONS system environment variable on the SHIR host is missing or configured with insufficient heap space. JVM heap space (Xmx) is below 8 GB, which is the minimum recommended for Parquet processing on ADF SHIR. Multiple Java installations on the SHIR host are conflicting. The SHIR host machine does not have the required JRE installed

How do I fix UserErrorParquetJavaInvocationException?

On the SHIR host machine, set the system environment variable: _JAVA_OPTIONS = -Xms256m -Xmx16g.. Restart all SHIR nodes after setting the environment variable.. Verify the correct JRE is installed (64-bit JRE 8 or later is required for Parquet processing on SHIR).. If multiple Java versions are installed, set the JAVA_HOME environment variable to point to the correct JRE.. After restarting the IR, rerun the pipeline to confirm the fix.

Medium severitygateway

Power BI Refresh Error:
UserErrorParquetJavaInvocationException

What does this error mean?

An Azure Data Factory copy activity writing or reading Parquet files failed because the JVM could not start or ran out of memory on the self-hosted integration runtime.

Common causes

1The _JAVA_OPTIONS system environment variable on the SHIR host is missing or configured with insufficient heap space
2JVM heap space (Xmx) is below 8 GB, which is the minimum recommended for Parquet processing on ADF SHIR
3Multiple Java installations on the SHIR host are conflicting
4The SHIR host machine does not have the required JRE installed

How to fix it

1On the SHIR host machine, set the system environment variable: _JAVA_OPTIONS = -Xms256m -Xmx16g.
2Restart all SHIR nodes after setting the environment variable.
3Verify the correct JRE is installed (64-bit JRE 8 or later is required for Parquet processing on SHIR).
4If multiple Java versions are installed, set the JAVA_HOME environment variable to point to the correct JRE.
5After restarting the IR, rerun the pipeline to confirm the fix.

Frequently asked questions

Does a self-hosted integration runtime failure affect all pipelines using it?

Yes — all pipelines routed through the affected SHIR fail simultaneously.

How do I check the health of a self-hosted integration runtime?

Open Microsoft Integration Runtime Configuration Manager on the SHIR host. The Status tab shows connection state and version. ADF Studio also shows SHIR health under Manage > Integration Runtimes.

What ports does a self-hosted integration runtime need outbound access to?

Outbound HTTPS (port 443) to *.servicebus.windows.net and *.core.windows.net. Individual connectors may need additional ports — SQL Server: 1433, FTP: 21.

Can I have multiple nodes in a self-hosted integration runtime for redundancy?

Yes — up to four nodes per SHIR. If one goes offline, ADF routes to a healthy node. Use at least two nodes for production pipelines.

Official documentation: https://learn.microsoft.com/en-us/azure/data-factory/format-parquet