docs(ci): starting a new failover runner needs svc.sh, not just config.sh

config.sh registers the instance without starting a listener, so the
procedure as written left the new runner offline and added no capacity.

.agents/notes/implemented/process/2026-07-26-ci-failover-runbook.i18n.yaml

@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write .agents/notes/implemented/process/2026-07-26-ci-failover-runbook.md
-2026-07-26-ci-failover-runbook.md: 05014454fa3e38045b89a857c346db0f897ab5a6
-2026-07-26-ci-failover-runbook.zh.md: e106a0de40799ca1c218217ea66c24068697dc53
+2026-07-26-ci-failover-runbook.md: b93c86d73f319f40706a4f9b31f448804e7b5ba8
+2026-07-26-ci-failover-runbook.zh.md: 25b83981e70070800c5a4037807d26811e55124d

.agents/notes/implemented/process/2026-07-26-ci-failover-runbook.md

@@ -28,7 +28,7 @@ Each of the three required Linux worker jobs — and the `all checks passed` ver
 
 ## Capacity during failover
 
-Six always-on instances absorb normal PR traffic (the pool's steady-state load is one serial standby job per master push, so failover capacity is effectively the full pool). If queues still build, register additional instances with an org registration token (org Settings → Actions → Runners → New runner). Clone an existing runner directory **excluding its identity files** — `rsync -a --exclude '.runner' --exclude '.credentials*' --exclude '_diag' --exclude '_work' <src>/ <dst>/` — then run `config.sh`; copying `.runner`/`.credentials` verbatim makes `config.sh` refuse with "already configured". About a minute per instance.
+Six always-on instances absorb normal PR traffic (the pool's steady-state load is one serial standby job per master push, so failover capacity is effectively the full pool). If queues still build, register additional instances with an org registration token (org Settings → Actions → Runners → New runner). Clone an existing runner directory **excluding its identity files** — `rsync -a --exclude '.runner' --exclude '.credentials*' --exclude '_diag' --exclude '_work' <src>/ <dst>/` — then run `config.sh`; copying `.runner`/`.credentials` verbatim makes `config.sh` refuse with "already configured". `config.sh` only registers the instance — it starts no listener, so a runner that stops there is registered and offline, adding no capacity. Install and start its service too: `sudo ./svc.sh install <user> && sudo ./svc.sh start` (this pool is systemd-managed; a foreground `./run.sh` also works but dies with the shell). Confirm the instance reports Idle in org Settings → Actions → Runners before counting it. About a minute per instance.
 
 
 ### Switch back

.agents/notes/implemented/process/2026-07-26-ci-failover-runbook.zh.md

@@ -28,7 +28,7 @@ Status: implemented
 
 ## 切换期间的容量
 
-6 个常驻实例可承接正常 PR 流量（该池平时唯一的稳态负载是每次 master 推送一个串行热备作业，故障切换时几乎全池可用）。若仍出现排队，用组织级注册 token（组织 Settings → Actions → Runners → New runner）追加注册实例。复制现有 runner 目录时**必须排除身份文件**——`rsync -a --exclude '.runner' --exclude '.credentials*' --exclude '_diag' --exclude '_work' <src>/ <dst>/`——再跑 `config.sh`；原样拷贝 `.runner`/`.credentials` 会使 `config.sh` 以 "already configured" 拒绝。每个约一分钟。
+6 个常驻实例可承接正常 PR 流量（该池平时唯一的稳态负载是每次 master 推送一个串行热备作业，故障切换时几乎全池可用）。若仍出现排队，用组织级注册 token（组织 Settings → Actions → Runners → New runner）追加注册实例。复制现有 runner 目录时**必须排除身份文件**——`rsync -a --exclude '.runner' --exclude '.credentials*' --exclude '_diag' --exclude '_work' <src>/ <dst>/`——再跑 `config.sh`；原样拷贝 `.runner`/`.credentials` 会使 `config.sh` 以 "already configured" 拒绝。`config.sh` 只完成注册，不启动监听进程，因此停在这一步的 runner 处于已注册但离线状态，不增加任何容量。还须安装并启动其服务：`sudo ./svc.sh install <user> && sudo ./svc.sh start`（本池由 systemd 管理；前台运行 `./run.sh` 亦可，但会随 shell 退出而终止）。确认该实例在组织 Settings → Actions → Runners 中显示 Idle 后再计入容量。每个约一分钟。
 
 
 ### 切回