How to run a Yellowfin cluster with session-replication on Kubernetes?
We are currently running a Yellowfin cluster on Google's Kubernetes environment. To make sure user session are working properly, we enabled sticky session (using a session cookie) on our Ingress.
In principle this works well, but whenever a pod gets restarted all current Yellowfin sessions pointing to that pod are killed. Some periods, the Kubernetes cluster is more volatile and we multiple restarts in a short timeframe. Unfortunately this leads to frustration with our stakeholders, as they face a lot of forced sign-offs.
I can recall that we tried to get session replication working on Kubernetes, but never succeeded.
We followed docs such as these and these.
Before giving this another go, I'd like to know if you are aware of any cases where the above did succeed? And if so: could you please point me to some doc's or other references?