Summary

Steps to Reproduce

1. Log in to the application
2. Create and configure a Cluster environment as follows:
   - Data Center that has 3 ignite servers and 6 Pega clients
   - Data Center two has 3 Ignite servers and  6 Pega clients
3. Reboot the failed Ignite server

Root Cause

A defect in Pegasystems’ code or rules. The service running on a platform is incorrectly reacting to standalone server nodes leaving the Cluster. When a node joins or leaves the cluster, the cluster notifies the interested parties of this event. 

When the platform is deployed in the client-server topology, where Pega acts as a client to a standalone Apache Ignite server, it does not filter the events. Services running on the client incorrectly react to servers joining or leaving the platform, assuming it is another client.

Resolution

1. Apply HFix-41862
2. Set the following setting to max values:

cluster/failure/detection/timeout

• valid range: (1second, 60 seconds)
• default: 10 seconds

cluster/network/timeout

• valid range: (500ms, 15 seconds)
• default: 5000ms

cluster/discovery/maxtimeout

• valid range: (1 minute, 15minutes)
• default: 10minutes