iterative / terraform-provider-iterative

☁️ Terraform plugin for machine learning workloads: spot instance recovery & auto-termination | AWS, GCP, Azure, Kubernetes
https://registry.terraform.io/providers/iterative/iterative/latest/docs
Apache License 2.0
290 stars 27 forks source link

task: generic option combinations not working #524

Closed casperdcl closed 2 years ago

casperdcl commented 2 years ago

Some useful generic option combinations don't work:

resource "iterative_task" ... {
  cloud   = "aws"
  image   = "nvidia"
  spot    = 0

  # willing to change this, need cheapest possible GPU:
  machine = "m+k80"
  region  = "us-east"
  ...
}
│ Error: operation error Auto Scaling: CreateAutoScalingGroup, https response error StatusCode: 
400, RequestID: 3c6bae3f-56f2-4a27-a0b5-e46b7323ae9b, api error ValidationError:
You must use a valid fully-formed launch template. The instance configuration for this
AWS Marketplace product is not supported. Please see the AWS Marketplace site
for more information about supported instance types, regions, and operating systems.

machine working:

not working:

0x2b3bfa0 commented 2 years ago

File a bug report with them, asking them to make the AMI available for all instance types. It should be in their best interest to allow that.

0x2b3bfa0 commented 2 years ago

More specifically, ask NVIDIA to enable the following instance types in the AWS Marketplace:

https://github.com/iterative/terraform-provider-iterative/blob/d4c34f8a21447e11af114a94b12499d3cac38e46/task/aws/resources/resource_launch_template.go#L61

https://github.com/iterative/terraform-provider-iterative/blob/d4c34f8a21447e11af114a94b12499d3cac38e46/task/aws/resources/resource_launch_template.go#L65

0x2b3bfa0 commented 2 years ago

at

casperdcl commented 2 years ago

Can we also have small instances for now https://github.com/iterative/terraform-provider-iterative/issues/328#issuecomment-1105637771?

0x2b3bfa0 commented 2 years ago

Action points for #524, i.e. this issue:

0x2b3bfa0 commented 2 years ago

Can we also have small instances for now (https://github.com/iterative/terraform-provider-iterative/issues/328)?

~Sounds like #328 to me. 🙃 Also known as p5-chore-we-actively-ignore unless we have a good reason not to.~ 😈

525