"...container-focused cluster management thing"
- For distributed data analytics using Dask in the cloud
- The basic workflow I'd like to master are
- Create a cluster in the cloud
- Using Dask, offload data preprocessing on cluster
- Scale up/down cluster according to needs
- The services I'll need to do this are
- Google Kubernetes Engine for creating the Kubernetes cluster
- Helm for package management in Kubernetes
- What lives inside a pod? Containers (e.g. Docker)
- How many containers can live in a pod? Any number, but usually just two
- Where do pods live? In a node
- What is a node? A physical or virtual machine (e.g. in GCP, a VM instance)
- How many pods can live in a node? Any number, but constrained by number of cores and memory available in the node
- A few things I've learned about Kubernetes by Julia Evans
- Contains good references for beginners
- See section on "Kubernetes from the ground up"
- Kubernetes BootCamp
- The Illustrated Childrens Guide to Kubernetes
- Managing Containers at Scale with CoreOS and Kubernetes by Kelsey Hightower