bruce48x/Lakona / .agents/skills/agar-three-node-e2e
agar-three-node-e2e
VerifiedRun the dedicated local three-node E2E smoke test for samples/Game.Unity.Agar through scripts/game/ci/test-agar-three-node.ps1. Use when the user asks to verify the Agar sample end to end, run the three-node topology, validate Docker Compose plus Unity PlayMode smoke behavior, or after changes to Game.Unity.Agar server/client code, cluster routing, gateway/battle/data node startup, Docker Compose configuration, or Unity multiplayer smoke tests.
Agar Three Node E2E
Overview
Use the existing repository script. Do not replace it with ad hoc Docker, Unity, or RPC commands unless you are debugging a specific failure after the script has produced artifacts.
The script starts the real samples/Game.Unity.Agar Docker Compose topology for
postgres, redis, data-1, gateway-1, and battle-1. It verifies the Redis
Membership Table contract and one shared Active view, runs Unity PlayMode
traffic against 127.0.0.1:20000/ws, restarts data-1 during the default live
traffic test, fences and replaces gateway-1, kills and recovers all three game
nodes, then reruns the default smoke after recovery.
Default Command
From the repository root:
pwsh -NoProfile -File scripts/game/ci/test-agar-three-node.ps1
Use PowerShell 7 or newer. The script itself has #Requires -Version 7.0.
Preconditions
Verify these before treating a failure as a product bug:
- Docker with
docker composeis installed and running. - Unity is installed, preferably the version in
samples/Game.Unity.Agar/Client/ProjectSettings/ProjectVersion.txt. - If Unity is not in the default Hub path, pass
-UnityPathor setUNITY_PATH. - Required ports are free or owned by the target test project's expected services.
- Run only one three-node test at a time, including across worktrees.
- The working tree can tolerate local artifacts under
.tmp/agar-three-node.
Useful Options
-UnityPath "<path-to-Unity.exe>": force the Unity executable.-TimeoutSeconds <seconds>: increase the total run deadline; must be at least60.-TestFilter <fully-qualified-test>: run another Unity PlayMode test. The default test opens the livedata-1restart window; another filter runs without that synchronized window but still performs gateway and whole-cluster recovery checks.-KeepEnvironment: leave Docker Compose services running after a successful or failed run.-ReuseEnvironment: retain the target project's containers and data instead of cleaning first; recovery tests still restart and replace nodes. Do not use against an environment serving other work.-SkipBuild: run Compose without--build; use only when the images are known current.-ProjectName <name>: select the Compose project; default islakona-agar-three-node-test. Ports, container names, and artifact paths remain fixed, so this does not enable parallel runs.
Examples:
pwsh -NoProfile -File scripts/game/ci/test-agar-three-node.ps1 -UnityPath "$env:ProgramFiles\Unity\Hub\Editor\2022.3.62f1\Editor\Unity.exe"
pwsh -NoProfile -File scripts/game/ci/test-agar-three-node.ps1 -TimeoutSeconds 900 -KeepEnvironment
pwsh -NoProfile -File scripts/game/ci/test-agar-three-node.ps1 -ReuseEnvironment -SkipBuild
pwsh -NoProfile -File scripts/game/ci/test-agar-three-node.ps1 -TestFilter "SampleClient.Gameplay.Tests.DotArenaTwentyClientLifecyclePlayModeTests.TwentyClientsCompleteMatchBattleSettlementAndLeaderboard" -TimeoutSeconds 600
Failure Triage
On failure, inspect .tmp/agar-three-node before rerunning:
docker-compose-startup.log: Compose startup command, stdout, stderr, timeout state, and exit code.docker-compose.ps.json: Compose service status snapshot.docker-compose.log: combined Compose logs.unity-editor.log: Unity batchmode editor log.TestResults.xml: Unity test result XML.lifecycle-report.json: node incarnation and recovery observations.
Use the failure phase to guide debugging:
- Preflight failure: check Unity path, Docker availability, or missing sample paths.
- Compose startup failure: inspect
docker-compose-startup.loganddocker-compose.log. - Membership contract failure: inspect the Redis service and the filtered
RedisMembershipTableTestsoutput before changing cluster runtime code. - Readiness or convergence timeout: inspect service health in
docker-compose.ps.json, node Membership views, and logs forpostgres,redis,data-1,gateway-1, andbattle-1. - Gateway port timeout: check
gateway-1endpoint binding and host port conflicts. - Unity test failure: inspect
unity-editor.logfirst, thenTestResults.xml. - Recovery failure: use
lifecycle-report.jsonanddocker-compose.logto determine whether an incarnation was reused, a Membership view failed to converge, or Actor Directory recovery rejected traffic.
Cleanup
The script normally runs:
docker compose -p lakona-agar-three-node-test -f samples/Game.Unity.Agar/docker-compose.yml -f .tmp/agar-three-node/docker-compose.local-test.override.yml down --volumes --remove-orphans
If -KeepEnvironment or -ReuseEnvironment was used, the script preserves the environment. Clean it manually with the same command when finished.
Keep this skill scoped to scripts/game/ci/test-agar-three-node.ps1; do not generalize it into a matrix runner or a replacement for the Godot chat E2E skill.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |