SA-21

Summary

ALL Agents stops running without any errors in logs. Agents have old timestamp for scheduled ‘Next Run’ indicating they were not triggered again. This article applies to PRPC running under Weblogic

Error Messages

None

Root Cause

- Agents have "Next Run" timestamp, indicating the last agent run completed successfully without any hang

- Enable DEBUG on class "com.pega.pegarules.session.internal.engineinterface.etier.impl.AsyncServant". Agent run is triggered asynchronously via JMS and DEBUG on AsyncServant class should write entry below if there are no issues with JMS configuration. Look at resolution notes if this DEBUG is not written

2014-06-16 15:43:10,655 [fault (self-tuning)'] [  STANDARD] [                    ] (       etier.impl.AsyncServant) DEBUG   - Received message to run agent task: false               AgentName           2              1402951390561  e72909b9257c28fd8f893f086100ff61 in Agent object 415715265

Resolution

Configure the JMS resources targeting EVERY server within the cluster

1. Create 3 JMS Servers targeting each server.
2. Create 3 JMS Module targeting each server.
3. Under each JMS Module create:
– a Connection Factory and subdeployment, targeted to the JMS Server
– a Topic and subdeployment, targeted to the JMS Server

4. Re-deploy the EAR to the cluster.
5. Stop-Start the cluster and monitor agent run.