Hello,

Background:

We are implementing Infrastructure as code cloud formation/Autoscaling scripts to build a Pega Environment.

We are using Redhat7 Linux AMI's for the same with parameterized context.xml, setenv.sh, and starting prweb as tomcat to bring up the node.

The question is as follows.

In our environment, we are using Kafka based Queue processors and Elastic search.

Since they are dependent on the stateful behavior, we have the following questions.

Search

1)If our environment has two search nodes, how is the replication maintained

2) When we have 2 search nodes, and one of them goes down and we spin up another node using IAC scripts/Autoscaling, do we have to re-index to populate the search index file or replication happens automatically? If we need to populate the index file, what's the recommendation.

3)When two of the search index hosting nodes go down, now since the complete index is lost, what is the recommended approach to reindex while spinning the node up using IAC scripts/Autoscaling? (PRPCUtils is what we target, but how can we populate the node information on the search landing page through command line)

Stream:

1)If our environment has two stream nodes, how is the replication maintained

2) When we have 2 stream nodes, and one of them goes down and we spin up another node using IAC scripts/Autoscaling, how is the master-slave relation maintained and how is the Kafka data be populated on the new node for the legacy events?

3)When two of the stream nodes goes down, How is the Kafka folder maintained if we spin up new nodes? How will this impact the queue processors which are scheduled?

Thanks,

Krishna