No description
  • TypeScript 45.2%
  • Go 28.4%
  • CSS 12.1%
  • HTML 9.4%
  • HCL 3.1%
  • Other 1.7%
Find a file
Derek Duncan 66f497de81 Run two runner instances and give the image a general toolchain
A runner process handles one job at a time, so concurrency comes from
running several on the host: runner_instances (default 2) unpack and
register core-infra-runner-1..N, each with its own work folder and its own
systemd service. Registration tokens are single use, so every instance
mints its own.

The image now carries what a general ARM64 workflow expects rather than
just what the Terraform job needed: build-essential, make, gcc,
pkg-config, nodejs, npm, golang-go, python3 tooling and rsync, plus
passwordless sudo for the runner user, which GitHub-hosted runners grant
and actions rely on. A 2 GiB swapfile keeps two concurrent jobs from
being OOM-killed on a 4 GiB host.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
2026-10-31 17:30:35 -05:00
.github Remove the temporary runner smoke workflow 2026-10-31 17:30:10 -05:00
apps update 2026-05-30 22:14:10 -05:00
broker updating 2026-05-09 16:22:01 -05:00
db update 2026-05-30 20:31:30 -05:00
gitea updating 2026-05-09 16:22:01 -05:00
observability update 2026-05-26 18:12:02 -04:00
scripts update 2026-04-25 22:56:58 -05:00
terraform Run two runner instances and give the image a general toolchain 2026-10-31 17:30:35 -05:00
.gitignore Add Infrastructure workflow, GCS state backend and a runner VM 2026-10-31 16:47:57 -05:00
Taskfile.yml updating taskfile 2026-09-05 00:33:42 -05:00
TODO.md large refector 2026-05-24 10:18:17 -05:00