AutoSys Performance Issues and Crashes after Migration to Oracle Exadata Cloud@Customer (ExaCC)
Article ID: 443918
Updated On:
Products
Issue/Introduction
Following a migration of the AutoSys (Workload Automation AE) database to Oracle Exadata Cloud@Customer (ExaCC), users may experience severe performance degradation or total application hangs.
The issue typically presents as a significant spike in memory usage by the AutoSys processes (specifically the event_demon), eventually leading to RAM and Swap exhaustion on the application server. During these episodes, the Scheduler may stop processing events, and administrators may find it impossible to connect to the AutoSys Manager without a hard reboot of the server.
System logs will often be flooded with performance warnings (CAUAJM_W_10998) indicating that query execution times have exceeded the configured DBLibThresholdTime, followed by unrecoverable query errors.
Environment
- Product: AutoSys Workload Automation AE
- Database: Oracle Exadata Cloud@Customer (ExaCC)
- Oracle Client Version: 19.22
Cause
A bug exists within the combination of Oracle Exadata Cloud@Customer (ExaCC) and Oracle Client version 19.22. This issue manifests as extreme memory consumption and database communication latency, causing the AutoSys application to starve for resources and eventually freeze or crash.
Resolution
Upgrading to version 19.31 resolves the memory exhaustion issue and restores stable performance to the AutoSys environment.
Additional Information
Feedback
