Hi everyone. I have a 4-node cluster of Horizon 2206 to be upgraded to 2312.1

When you guys upgrade multi-node clusters do you just schedule a total down time or do you just use the disable function and do one server at a time along with getting your network engineers to disable the node as a load balancer target?

Thanks very much. 

Hi @MattDsa 

In our environment, we have two pods that allow us to implement a staggered update approach for Horizon releases. When updating to a new release, we initially switch the load balancer (LB) to the secondary site. The following morning, we conduct internal testing to verify that everything is functioning as expected.

After a period of one to two weeks, we proceed with updating the next pod. This phased approach provides an added layer of safety, as it enables us to roll back if necessary and also allows us to obtain confirmation from users that the update has been successful.

We do a rolling update by disabeling one connection server at a time. Our load balancing detects the disabeling so we do not need to do anything there.