Active/Passive Cloud hosted Kubernetes support
We need to know if Pega supports an active/passive Cloud hosted Kubernetes CDH implementation using Postgres as the RDBMS on version 8.8.x. If so, whom can we work with to identify the recommended architecture and configuration, including replication of all layers including externalized services (i.e. Cassandra, Kafka, and SRS).
Note: We have set up an active/passive multi-datacenter configuration on Pega 7.4 using VM's, Oracle RAC, and GoldenGate. However, we are unsure how to do this with Pega 8.8.x running in Kubernetes with Postgres, and now that Cassandra, Kafka, and SRS are externalized.
Pegasystems Inc.
US
@SriHarsha Anika We have Pega deployed on a client-managed cloud. We would like to set up a passive deployment in separate region for failover and disaster recovery.
Pegasystems Inc.
IN
Hi My Question is client has deployed Pega in AWS/AZURE/GKS what sort of cloud it is?
For your requirement you should leverage database service of respective cloud vendor which will take care of cloud back up and data synch between active and passive database spread across different datacenters.
Similar is the case for the application server PODS.
Cloud vendor will have the required resilience service which will turn on your passive clusters and pods on the event of failure. you need to learn , work with cloud vendor to understand the feature and configure it.
It is not a apple to apple match like an on-premise configuration in cloud. In Cloud everything is a service and a configuration.
Pegasystems Inc.
US
Hi. We run our Pega implementation on AWS. I don't believe the statement "In Cloud everything is a service and a configuration." is accurate. For example, we are running Cassandra on EC2's as I was told that AWS Keyspaces is not supported as DDS nodes.
Also, AWS will replicate Postgres from one region to another; however, the replica is a full read only database, which may be suitable for DR but not for active/passive. In our on-premise multi-datacenter implementation, we only replicate transactional data (e.g. IH tables) from the RDS that would be needed for business continuity in the event of failover.
Also, what considerations would be needed for CDH, ADM, Stream nodes, and SRS?
Thanks.