Is HashiCorp Nomad a Better Kubernetes?(chaordic.io)
chaordic.io
Is HashiCorp Nomad a Better Kubernetes?
https://chaordic.io/blog/is-nomad-a-better-kubernetes/
6 comments
Yes, it is. Kubernetes is insanely complicated. Nomad is industrial grade but still simple.
This article displays an astounding level of ignorance about how to run workloads absent a container orchestrator.
If you run a docker container on autoscaling EC2 instances that register with a target group, you have apparently "written your own orchestrator."
If you run a docker container on autoscaling EC2 instances that register with a target group, you have apparently "written your own orchestrator."
How did you arrive at that conclusion? The post claims nothing of the sort.
The post provides a laundry list of things that a container orchestrator does for you, and asserts that if you don't use Kubernetes or Nomad you have to implement all those things yourself.
> If you have a set of servers, but no scheduler like Nomad or Kubernetes, you would be forced to do the following on your own:
> Find the optimal server to place the execution of software on, for each deployment, given existing resource utilization of other processes & hardware.
> ...
If you have autoscaling groups there isn't any scheduler that I have to worry about (Amazon of course has one). Certainly I don't have to reinvent ten wheels just because I choose not to use a container orchestrator.
> If you have a set of servers, but no scheduler like Nomad or Kubernetes, you would be forced to do the following on your own:
> Find the optimal server to place the execution of software on, for each deployment, given existing resource utilization of other processes & hardware.
> ...
If you have autoscaling groups there isn't any scheduler that I have to worry about (Amazon of course has one). Certainly I don't have to reinvent ten wheels just because I choose not to use a container orchestrator.
Autoscaling groups solve the problem of ensuring you have the desired number of server nodes running.
They do not in anyway solve optimal placement of jobs. It only works “optimally” if you are running just a single service, or services that can all run fine on a single node, in which case you don’t need a scheduler to begin with. The Autoscaling group in this case only provides you with scaling & redundancy, not scheduling capabilities.
> Any headline that ends in a question mark can be answered by the word no.