Edit

‎Resolve Worker out-of-memory failures

In certain cases, for instance, when you are processing large-size files, Workers can fail due to an out-of-memory error. A potential solution to address these errors is to increase the default Worker memory setting and enable automatic Heap Dump generation.

The article contains instructions on increasing the default memory settings and enabling Heap Dump generation for Control Tower (CT) and RPA Workers.

CT Workers

To increase the memory and tune Heap Dump generation for CT Workers, follow the steps below:

1. Log in to the BEP Master server via SSH.

2. Go to $INSTALL_DIR/vds-data/workers/app/com.workfusion.spa.ct/worker-app/<version>/ and back up the worker.yml file.

3. In the worker.yml file, increase the value for the heap parameter from 1024 to 2048 or more.

4. After -XX:+ExitOnOutOfMemoryError, add the following parameters:

   • XX:HeapDumpOnOutOfMemoryError
   • XX:HeapDumpPath=/opt/workfusion/heap_dumps/heapdump.hprof
   • Xloggc:/opt/workfusion/heap_dumps/gc.log
   • XX:-PrintGCDetails
   • verbose:gc
   • XX:+PrintGCDateStamps

   Note that the values for XX:HeapDumpPath and Xloggc can be different from those in the example above, for instance: -XX:HeapDumpPath=/opt/heapdump.hprof and Xloggc:/opt/gc.log.

5. In Marathon UI, destroy CT Worker jobs. After that, new Workers start with modified parameters.

   

RPA Workers

To increase the memory and tune Heap Dump generation for RPA Workers, follow the steps below:

1. In RPA\bot-agent\conf\bot-agent-unit.yml on the RPA server, increase the following parameters:

   • Xms

   • Xmx

   • XX:MaxMetaspaceSize=

2. For heap dump generation, add the following parameters:

   • XX:+HeapDumpOnOutOfMemoryError
   • XX:HeapDumpPath=../logs/unit${unit_id}

3. Restart the Bot Unit with the Worker. To do that, navigate to Task Manager > Users, right-click the corresponding Bot Unit, and choose Sign Off. After that, the RDP window restarts.