I haven't checked the load balancer logs. But let me know whether this error messages causes the cases to be lost i.e. work object is failed to save into DB.
One thing I want to highlight here is that case is getting lost for the request coming from other channel(WEB). But when we run the same request manually, it is successfully run and created the case accordingly.
Not yet. We use SAML2.0 Authentication and Auth Service for login. When we run one 1 node it works fine. When we move to 2 nodes it fails 8/10 times, so inconsistent. We are on AWS Public cloud with Application Load Balancer (ALB). When it fails I see this message in the logs:
2021-02-24 14:28:44,276 [otejob-executor-4897] [ STANDARD] [ ] [ ] ( internal.mgmt.PRRequestorImpl) INFO - [QuiescePassivation] forcePassivateInner: passivated or destroyed requestor [H2M094QKEOAARWXTEHFM6UL495E884D4FA]
So yes as Ryan commented it is a HC or Passivation setting issue which kicks in when 2 nodes are in play.
Posted: 2 months ago
Updated: 2 months ago
Posted: 3 Feb 2021 15:14 EST Updated: 3 Feb 2021 16:09 EST
Quiessence passivation occurs when the node is gracefully shutting down. This means there is an issue on your JVM and the healthcheck response is coming back too slow. Pega receives a shutdown command and starts the process to passivate your requestors. This is either an issue with your jvm stability or your passivation settings.