Handling PongMissed Errors in Opcenter Execution Pharma Services

Problem

PongMissed errors occur when there are network inefficiencies or temporary disconnections between the database and the application server. This can lead to synchronization issues in the services. Although Oracle alert logs (ORA errors) might indicate these issues.

some Oracle error examples:

• ORA-12608: TNS: Send timeout occurred
• TNS-12637: Packet receive failed
• ORA-12541: TNS:no listene
• ...

Solution

To resolve PongMissed errors, consider both infrastructure verification and MES application configuration changes:

1. Infrastructure Check: Ensure there are no database or network inefficiencies causing disconnections or slowdowns.

2. Adjust Synchronization Parameters: Update the following XFP_PARAM synchronization settings on the Opcenter side to allow the service controller to better manage disconnections:

• SERVER_SYNCHRO_DELAY:
  • Description: Delay between two synchronizations of an Opcenter Execution Pharma service with its Controller.
  • Action: Change from 10 seconds (default) to 30 seconds.

• SERVER_SYNCHROTIMEOUT:
  • Description: Service start and stop timeout duration.
  • Action: Change from 30 seconds (default) to 90 seconds.

• SERVER_MAX_START_RETRY:
  • Description: Maximum number of attempts to restart a service before setting its status to Failed.
  • Action: Change from 5 attempts (default) to 10 attempts.

3. Restart Requirement: After modifying these parameters, remember to restart the Opcenter Execution Pharma Controller Service to apply the changes.

4. Task Timeout Adjustment (PI Services):
   • If tasks in PI services are timing out (taking too long), extend the timeout using:
   • PISERVICE_EXECUTION_TIMEOUT: Change from the default value of 15 seconds to 30 seconds.

Implementing these configurations can help mitigate PongMissed incidents by increasing the resilience of service synchronization against temporary network issues.