I have a Rancher Sever deployed on AWS using ECS and auto-scaling groups. This works great however I’m creating a second ECS cluster for my Rancher Hosts in which I have an ECS task definition for the rancher/agent container, this is getting scheduled but because it only runs for a short period of time ECS keeps re-scheduling each time it stops causing multiple rancher/agents containers on the hosts. Does anyone know a way round this or should I be using an alternative method all together for running the agent?
Apologies if this is a very stupid question, but can rancher allocate containers on ECS? Or is there a guide for running the current “most used config” (i.e. k8s) on AWS using ECS only or ECS/EC2 combination?
My understanding from the tutorials is the Rancher server runs as a container service straight on ECS, but the rancher administered hosts are still running under EC2, is that correct? Does anyone have a link to a tutorial which is still valid? There are a few screencasts and tutorials but I cant seem to pinpoint which is the “current” one?