runs-on, once the GitHub App is installed and your
organization is turned on.
1. Install the GitHub App
Install the GitHub App on your organization and choose the repositories. During the beta, we turn your organization on by hand and send you its label; jobs wait until then. See the quickstart for each step.2. Change runs-on
3. Map the sizes
New organizations start at 2 vCPU, so start with your 2 vCPU jobs:blacksmith-2vcpu-ubuntu-2404andblacksmith-2vcpu-ubuntu-2604map tocoreload-2.- Blacksmith’s 4 and 8 vCPU jobs: move each to a Coreload size your organization has (the table below).
blacksmith-16vcpu-*has no Coreload size today: keep those jobs on Blacksmith.blacksmith-32vcpu-*has no Coreload size: keep those jobs on Blacksmith.- Blacksmith’s ARM (
-arm), Windows and Ubuntu 22.04 labels have no Coreload equivalent: keep those jobs on Blacksmith.
Billing starts when prepaid credit opens; until then nothing is charged.
At 2 and 4 vCPU, a Coreload job has less memory than a Blacksmith job of the same size (see the table). Keep a job that runs
close to its memory limit on Blacksmith for now.
4. Check what differs today
- The image. Coreload jobs run on a Debian 12 image with Node.js, npm, Python 3 and git, not on an image like
GitHub’s Ubuntu runners.
actions/setup-pythoncan fail to download a Python build, and Go is not preinstalled. - Docker. There is no Docker in a Coreload job yet. Keep jobs that build or run containers on Blacksmith for now,
including those that use
useblacksmith/setup-docker-builderoruseblacksmith/build-push-action. - Caching. A job’s artifact and cache calls may go to a Coreload relay on the machine, or straight to GitHub. Through the relay, artifact calls go on to GitHub with the job’s own token, and cache calls go to a Coreload service on the machine without it and miss, since there is no Coreload cache yet.
- Sticky disks and SSH. Blacksmith’s sticky disks and SSH access have no Coreload equivalent yet.