I'm very new to KOPS, so I have very questions about the cluster deployment process in AWS.
Supose I have 3 reserved instances: I want to deploy one Master and two workers in my existing AWS reserved instances.
Can anyone write a step by step guide to acomplsih that?
Thanks.
Hi @yngwuoso, you shouldn't need to do anything special for reserved instances. Simply follow a guide, but set your master and node size to your reserved instance sizes.
https://github.com/kubernetes/kops/blob/master/docs/aws.md
Add these flags to kops create cluster
--master-size m3.medium
--node-size m3.large
Hi Mike. Doesn't Kops create on demand instances in the deployment process? I say that, because when I purchase reserved instances, they are present at the "Reserved Instances" section at AWS Console. I thougth that, I have to configure that particular instances somewhere in the configuration files in Kops to use them (avoiding the creation of other instances).
Ok, reserved instances are a "billing contract", not a virtual machine... So, you are right.
Thanks for the information.
Hey Guys! Sorry to bring this issue back but I have a question about this topic.
As I understood from the comments here, there is no such thing as a reserved instance in practice. As @Yngwuoso said previously, _reserved instances are a billing contract, not a virtual machine_.
My question is: supposing I already have a cluster running, if I buy a reserved instance, with the same size as one of the nodes on my cluster, that node will be priced the reserved instances from that moment? If not, then I can destroy that node and kops will bring a new instance, this one as a "reserved instance"?
Nevermind. I already got the answer.
Basically reserved instances are as a promotional code for your billing. You don't need to do anything more than buying the reserved instances in the proper zone and with the proper size.
Most helpful comment
Ok, reserved instances are a "billing contract", not a virtual machine... So, you are right.
Thanks for the information.