Blog outlines setup to provide virtual hostname to HANA scale-out environment
via load balancing on google cloud. so even in case of failed-over to standby , application connections will divert automatically to available node.in short NO OUTAGE for applications due to failed-over scenario.
Google cloud load balancing use case
As per standard setup HANA scale-out comes with MASTER , WORKERS and standby role on severs. Application (SLT , BIA tools etc.) connects to one of MASTER OR WORKER node. but application connection fails once that node crashes and failed-over to standby. as per current scenario manual effort require to change all clients connection properties which is time consuming.
In case of SLT data transfer - during outage backlog increases and it might also impacts BIA reporting i.e. true business impact.
Cloud load balancing adds healthcheck feature which keep on checking HANA tenant index server ports and diverts connections only to active nodes. it keeps application connectivity healthy.
HANA statement routing - Concept behind load balancing
parameter client_distribution_mode --> statement
Configuration at google cloud layer
Configuration will add below components at google cloud level. they can be create in sequence as per diagram via command line OR directly from cloud console. ensure you have access at cloud console level.
Post-setup details at google cloud console
After setup health check will look like below at cloud console. it means group ig1 has standby node status where HANA index server and respective port is not open. so application connections will NOT go there.
Health check status will change if failed-over occurs and application connections goes according to active nodes.
Once setup is done configure virtual hostname with application connection.