Pega 81 and Kafka
Stream nodes spawns the kafka brokers jvms. Replication happens between Kaka brokers.Pega uses Charlatan ( instead of zookeeper) and stores info in Pega DB. Kafa Brokers keeps data in filesystem.
If JVM communication between datacenter is not allowed ( as in our case), we will have to disable Kafka. The side effects of disabling it will be that we will not be able to use Queue Processors and stick to Standard Agents.
Below are the steps we need to follow to disable Kafka.
- Go to Decisioning → Infrastructure → Services → Stream
2. Decommission Stream Nodes
3. Remove the stream nodes from DB
TRUNCATE TABLE PRPCDBA.PR_DATA_STREAM_NODES;
TRUNCATE TABLE PRPCDBA.PR_DATA_STREAM_NODE_UPDATES;
TRUNCATE TABLE PRPCDBA.PR_DATA_STREAM_SESSIONS;
4. Set the below DSS to false.
5. Ensure none of the nodes have Stream node type.
-DNodeType=Stream
JVM restart is required after these changes.
Senior Professional Manager - DXC Technology
3 年very Useful
Pega Developer at ABC
4 年Nice post. thank you. Just some quick questions, if you disable the embedded kafka and use external kafka instead, will the charlatan still be used? Will the events/topics still be saved in the Pega database or just in the external kafka or both?
Vice President at Morgan Stanley
5 年Do you see any issues if stream nodes are enabled?