Spare Cores Runner is a command-line tool and Python API designed to simplify the process of provisioning and managing cloud instances across various cloud providers. It leverages Pulumi to handle infrastructure-as-code and automate the creation, and destruction of compute instances in your preferred cloud environment.
Spare Cores Runner (sc-runner
from now on) is used by Spare Cores Inspector
to start up basic cloud instances with the required minimum configuration and environment to start the machine with a custom
cloud-init
script, but it can be used to just start an instance with a given SSH key, so it's available for basic
uses.
sc-runner
uses the Spare Cores database to pre-fill the available configuration options, like what regions and instance
types are available for a given cloud provider.
To use a given cloud provider, you have to provide credentials for them, either by running it in an environment,
where the underlying library can pick them up, or by specifying them through environment variables or by command line options.
For more details, see the supported vendor's Pulumi integration:
As sc-runner
uses Pulumi under the hood to manage the cloud providers, it'll need to have a Pulumi project.
You can configure Pulumi project-related settings either by specifying them using command line arguments, shown below:
sc-runner create --help
Usage: sc-runner create [OPTIONS] COMMAND [ARGS]...
Options:
--project-name TEXT Pulumi project name [default: runner]
--work-dir TEXT Pulumi work dir [default: /data/workdir]
--pulumi-home TEXT Pulumi home [default: /root/.pulumi]
--pulumi-backend-url TEXT Pulumi backend URL [default: file:///data/backend]
--stack-name TEXT Pulumi stack name, defaults to
{vendor}.{region}.{zone}.{instance_id} or similar
--help Show this message and exit.
Commands:
aws
azure
gcp
hcloud
upcloud
Or by setting the following environment variables:
- PULUMI_PROJECT_NAME
- PULUMI_WORK_DIR
- PULUMI_HOMEPULUMI_BACKEND_URL
sc-runner
should create the project on the first invocation, in which it'll create multiple stacks for each
vendor.region.zone.instance_id
tuple. This is to allow concurrent creation of the instances, supporting our
Spare Cores Inspector use case, where we start a given instance type once for collection data from them.
Create and destroy a default t3.micro
instance in the default VPC/subnet in AWS:
sc-runner create aws
sc-runner destroy aws
Create an Azure instance with a public key:
sc-runner create azure --instance Standard_DS1_v2 --public-key $(cat .ssh/id_ed25519.pub)
Create an AWS instance with public IP and an already stored SSH key named spare_cores
:
sc-runner create aws --region us-west-2 --instance t4g.large --instance-opts '{"associate_public_ip_address": true,"key_name":"spare-cores"}' --public-key ""
sc-runner destroy azure --instance Standard_DS1_v2
sc-runner destroy aws --region us-west-2 --instance t4g.large
You can also user destroy-stack
instead of destroy
, which does a pulumi refresh
first, synchronizing the underlying
cloud-reality with Pulumi's internal backend and destroys only what's really there (so it won't fail on already deleted
resources).
Sometimes Pulumi might leave a lock file in its state store, preventing further operations. With this command you can cancel those:
sc-runner cancel aws --region us-west-2 --instance t4g.large
sc-runner
is available through a Docker image as well, which you can use with the following command:
docker run --rm -ti ghcr.io/sparecores/sc-runner:main --help
You'll have to set up the same environment variables or configure the cloud credentials for Pulumi.