ci: full-parity and PR parity jobs on playhorizon, not the desktop

The maintainer asked that CI stop using his desktop (2026-10-26). playhorizon's
runner now carries horizon-game and the desktop's runner none of the job labels;
full-parity jobs go to playhorizon only, with a 6 h timeout on its 4 CPUs.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Ludovic Marie
2026-10-26 23:46:29 +01:00
parent a727f617f0
commit 111db2827b
3 changed files with 25 additions and 20 deletions
+2 -2
View File
@@ -25,8 +25,8 @@ jobs:
parity-main:
# A Mac mini runner of its own (ludovics-mac-mini-2 holds only this label since 2026-10-26): behind PR parity jobs
# this run waited 2 h while every merge replaced the pending one, and on the release runner a release then waited
# behind the next run. PR parity jobs have ludovics-mac-mini and the desktop. Not the desktop: a job slowed down
# because the maintainer came back could time out, and a failure here bisects and comments on merges.
# behind the next run. PR parity jobs have ludovics-mac-mini and playhorizon. Not playhorizon: its 4 CPUs
# make a core-crate run several times longer, and a failure here bisects and comments on merges.
runs-on: [self-hosted, macOS, horizon-parity-main]
timeout-minutes: 240 # a core-crate batch reruns ~300 harnesses (~60-90 min), a bisect a few rounds of the failures
steps:
+8 -8
View File
@@ -1,6 +1,6 @@
name: parity
# The vanilla harnesses a pull request's changes can affect, on the Mac mini's runner, which holds the maintainer's
# own Minecraft files (Horizon never downloads the game). Called by ci.yml for pull requests from this repository
# The vanilla harnesses a pull request's changes can affect, on the Mac mini's runner or playhorizon's, which hold the
# maintainer's own Minecraft files (Horizon never downloads the game). Called by ci.yml for pull requests from this repository
# and by run-parity.yml when a maintainer labels a fork's pull request run-parity. See docs/release-flow.md.
on:
workflow_call:
@@ -16,8 +16,8 @@ on:
full:
description: >-
The pull request carries the full-parity label (a change every harness reaches, such as a crate split): it
runs on the desktop's runner only, so the Mac mini's runners stay free for the others. While the desktop is
in use the job waits for it.
runs on playhorizon's runner only, so the Mac mini's runners stay free for the others. While that runner is
busy the job waits for it.
type: boolean
default: false
permissions:
@@ -28,10 +28,10 @@ env:
jobs:
parity:
runs-on: ${{ inputs.full && fromJSON('["self-hosted", "horizon-game", "ludodesktop"]') || fromJSON('["self-hosted", "horizon-game"]') }}
# A change to core crates reruns ~350 harnesses: ~60-90 min on the Mac mini, ~35 min on the desktop, longer if
# someone comes back to the desktop meanwhile (the job then goes on with 2 threads).
timeout-minutes: ${{ inputs.full && 300 || 120 }}
runs-on: ${{ inputs.full && fromJSON('["self-hosted", "horizon-game", "playhorizon"]') || fromJSON('["self-hosted", "horizon-game"]') }}
# A change to core crates reruns ~350 harnesses: ~60-90 min on the Mac mini; on playhorizon (4 of its 8 vCPUs, 4
# harnesses at a time) up to a few hours, and any pull request's job may land there.
timeout-minutes: ${{ inputs.full && 360 || 180 }}
steps:
- uses: actions/checkout@v4
with:
+15 -10
View File
@@ -13,7 +13,7 @@ reopened; labels never start it):
| Job | Runs on | What | Time |
| --- | --- | --- | --- |
| `verify` | GitHub `ubuntu-latest` | Public-tree check, Python tests, rustfmt, clippy with warnings denied, Rust tests. Dependencies and the workspace build come from a cache saved by main's runs | ~4 min cold, less warm |
| `parity / parity` | A runner with the label `horizon-game`: the Mac mini's two, and the desktop's while nobody uses it (below) | Only for a change to a harness, an oracle, a probe or the runner (or the `run-parity` label): the vanilla harnesses the change can affect (`tools/parity.py --changed <base>`). Other pull requests skip it: [parity-main](#parity-on-main) checks them on main | seconds when nothing is affected; minutes per affected harness |
| `parity / parity` | A runner with the label `horizon-game`: the Mac mini's and playhorizon's (below) | Only for a change to a harness, an oracle, a probe or the runner (or the `run-parity` label): the vanilla harnesses the change can affect (`tools/parity.py --changed <base>`). Other pull requests skip it: [parity-main](#parity-on-main) checks them on main | seconds when nothing is affected; minutes per affected harness |
Pushes to main run `verify` only; the nightly audit covers main's harnesses.
@@ -29,12 +29,12 @@ repository, `run-parity.yml` for a fork's pull request a maintainer labels `run-
A pull request every harness reaches (a crate split moves files every probe links: #2160 ran all 352 harnesses, 71
minutes on a Mac mini runner, while five other pull requests' jobs queued behind it) carries the `full-parity` label:
its job runs on the desktop's runner only (8 harnesses at a time), and waits for it while the desktop is in use
instead of taking a Mac mini runner. Labels don't start `checks`: add it when opening the pull request
its job runs on playhorizon's runner only (4 harnesses at a time), and waits for it while that runner is busy
instead of taking the Mac mini's. Labels don't start `checks`: add it when opening the pull request
(`gh pr create --label full-parity`), or before the next push.
Horizon never downloads the game, so harnesses that compare with vanilla run only where the maintainer's own
copy is: the Mac mini, his Linux desktop and playhorizon (below). Their runners run as the maintainer's account and keep their checkout
copy is: the Mac mini and playhorizon (below). Their runners run as the maintainer's account and keep their checkout
between runs (`clean: false`), so the probes build incrementally into `.local/parity-target` and earlier passes
stand.
@@ -83,8 +83,9 @@ release's macOS build back. Each keeps its own checkout, `target/` and Rust inst
`_work/horizon/shared-local` links to the first's, so they share the vanilla data, the preparations (one at a
time, under a lock) and the parity passes and slots.
**The desktop's runner.** `ludodesktop-horizon` (labels `self-hosted`, `Linux`, `X64`, `ludodesktop`, and
`horizon-game` once a full trial has passed there) runs on the maintainer's Linux desktop while nobody uses it: its
**The desktop's runner (retired).** On 2026-10-26 the maintainer asked that CI stop using his desktop: its services are
stopped and `ludodesktop-horizon` has no label a job asks for (only `self-hosted`, `Linux`, `X64`), so nothing waits for
it. What follows describes it as it ran. It ran on the maintainer's Linux desktop while nobody used it: its
screen locked, or untouched for 15 minutes (his rule, 2026-10-26). `tools/ci/linux-runner.sh install`, run there as
him, registers it and installs two systemd user services: `horizon-runner` (the runner, in
`~/.actions-runners/horizon`, nice 19, in `horizon-ci.slice` with CPU and I/O weight 20, half the threads (on all of
@@ -105,12 +106,16 @@ and its failures bisect and comment on merges.
**playhorizon's runner.** `playhorizon-horizon` runs on a server a hosting partner lends the project (Ubuntu, 8
AMD EPYC threads, 30 GB, in Detroit; reachable only through the maintainer's tailnet). Nobody sits at it, so the
watcher never pauses it except by hand, and it gets the machine: `RUNNER_CPUS=7 RUNNER_MEMORY=24 RUNNER_JOBS=6`.
watcher never pauses it except by hand. It gets half the machine: `RUNNER_MEMORY=24`, `CARGO_BUILD_JOBS` and the
harness slots at 4, and a drop-in (`~/.config/systemd/user/horizon-runner.service.d/cpus.conf`) that starts it under
`taskset -c 4-7`, so CPUs 0-3 stay free for benchmarks pinned there (the user slice can't set `AllowedCPUs`: systemd
doesn't delegate the cpuset controller to it).
Without Nix, `linux-runner.sh` takes the build libraries, the JDK, sccache, Xvfb and Mesa's software Vulkan
(lavapipe; there is no GPU) from apt, and Python 3.13 from uv. The partner can reach the machine, so gh is not
logged in there: the registration token comes from another machine (`RUNNER_TOKEN`), and no job with secrets may
run there (parity has none). It is also a parity host (`playhorizon` in `tools/parity_hosts.json`, 6 jobs). Like
the desktop's, it gets the `horizon-game` label only once a full trial has passed there.
run there (parity has none). It is also a parity host (`playhorizon` in `tools/parity_hosts.json`, 6 jobs). Since
2026-10-26 it has the `horizon-game` label: pull requests' parity jobs run there or on the Mac mini, and `full-parity`
ones only there.
## Merging
@@ -218,7 +223,7 @@ How releases are numbered, and how installs from before 0.2 reach it: [versionin
| `release-now` | Releases as soon as main's checks pass, without waiting for merges to settle |
| `skip-release` | Doesn't trigger a release and isn't listed in the notes |
| `run-parity` | Runs the parity job once on a fork's pull request |
| `full-parity` | Runs the pull request's parity job on the desktop's runner only (a change every harness reaches) |
| `full-parity` | Runs the pull request's parity job on playhorizon's runner only (a change every harness reaches) |
| `parity-regression` | Opened by the nightly audit for a harness that fails on main |
| `release-blocker` | On an issue: no game release is published while it is open ([One channel](#one-channel)) |