Workato · On-prem agent
Failed to initialize MemoryUtil (Databricks JDBC)
Short answer
This error comes from the Databricks JDBC driver's use of Apache Arrow for data transfer. Add EnableArrow=0 to the JDBC URL.
VerifiedLast verified
01Error text
java.lang.ExceptionInInitializerError: Exception java.lang.RuntimeException: Failed to initialize MemoryUtil. [in thread "pool-21-thread-1"]02Likely causes
- The Databricks driver uses Apache Arrow for data transfer.
03How to fix it
- Add
EnableArrow=0to the JDBC URL, for examplejdbc:spark://host.cloud.databricks.com:443/default;EnableArrow=0;. - Restart the agent and rerun the job.
04Official documentation
→Related errors
- Snowflake JDBC driver internal error (ArrowBuf.getDirectBuffer)This error appears on a select query in a custom SQL action against Snowflake through the on-prem agent. Add &JDBC_QUERY_RESULT_FORMAT=JSON to the connection URL.
- java.lang.OutOfMemoryError: Java heap space (on-prem agent)This error in the on-prem agent log means the agent does not have enough memory for a large job or payload. Raise the Java heap with the -Xmx option and restart the agent.
Test every fix in a non-production environment before you apply it to production. See how entries are verified.
More for Workato work:Open Workato jobsWorkato rates and salariesHire a senior Workato developer