Considerations:
Questions:
Phases:
| zero | multi-node Kind (Docker) cluster |
| one | multi-node physical cluster |
| two | + high availability zones |
| three | + netbooting nodes |
| four | multiple multi-node physical clusters |
Clusters:
I am keeping the hardware in each cluster as homogeneous as possible just to keep the infrastructure management simpler. I realize this will give me less opportunity to “train” on heterogeneous deployments (like at my present employer) but it is always best when you have the option. In fact, I want to train myself in working with the best solutions at home, not reproducing the cluster-fucks at different places of employment.