Question
CBA
AU
Last activity: 23 Apr 2020 15:06 EDT
stream service is not enabled by default when new node provisioned
Our application is running on AWS cloud, whenever a node goes down for any failure, new instance is provisioned by auto heal/scale functionality. but the stream service is not in normal state by default in the new node. Its in joining failed status, since the previous node is in unreachable state. this create lot of exceptions from queue processors. for now we are manually decommissioning the old node and start the new node. can we do this programmatically ? whenever new node comes we need to remove the unreachable nodes automatically
Pegasystems Inc.
IN
Hi Prabhu,
This happens because if the node got terminated ,still the stale nod entries would present in pr_sys_statusnodes .
Cluster might not have restarted which will delete all the node entries in statusnodes table and come up with new nodes.
However,by saying this for multi-node environments cluster restart for single node termination would be costly operation .
we have been seeing this issue for multiple customers and I think you can raise a SR on this so that we can investigate this behavior further.
Thank you,
Arun
CBA
AU
Hi Arun,
After truncating pr_data_stream_nodes table, the performance was much better even stream service not enabled.
May I know the significance of this table pr_data_stream_nodes? and at what circumstances should we truncate this table.
Thanks,
Prabhu
Pegasystems Inc.
IN
You can go through below article for pr_data_stream_nodes and when you can truncate
https://community.pega.com/knowledgebase/articles/decision-management-overview/kafka-operations
https://community.pega.com/knowledgebase/articles/decision-management-overview/troubleshooting-kafka