feat: 推送取消 — 引擎 cancel flag 检查 + WS cancelled 计数器

sync_engine_v2: 每台服务器推送前检查 Redis sync:cancel:{batch_id}，
若已取消则跳过并标记 cancelled 状态，WS 广播 cancelled 计数。
websocket: broadcast_sync_progress 新增 cancelled 参数。

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

docs/design/specs/2026-05-30-terminal-quick-cmd-design.md

@@ -0,0 +1,56 @@
+# 2026-05-30 终端页面快速命令栏 — 设计文档
+
+## 背景与目标
+
+terminal.html 当前只能手动输入命令，运维高频操作（如查看服务状态、读日志、重启进程）每次都要手敲。需要一个**可自定义的快速命令栏**，一键发送预设命令到终端。
+
+## 方案
+
+| 方案 | 优点 | 缺点 |
+|------|------|------|
+| A: 右侧面板底部固定命令栏 | 不占屏幕空间，可折叠 | 命令多了需要滚动 |
+| B: 工具栏下方横排按钮 | 一眼可见 | 横向空间有限（2-3个） |
+| C: 右键菜单 + 自定义命令 | 上下文触发 | 交互多一步 |
+
+**选定 A + C 混合** — 右侧面板加"快速命令"区域，同时支持自定义（添加/编辑/删除），存储在 `localStorage`。
+
+## 命令来源
+
+| 来源 | 说明 |
+|------|------|
+| 内置默认 | 5 个通用运维命令：`systemctl status` / `journalctl -f` / `df -h` / `free -h` / `top -bn1` |
+| 用户自定义 | 点击 + 号添加，输入命令名 + 命令内容，存 localStorage |
+| 会话记录 | 不自动添加（避免混乱，用户手动保存常用命令） |
+
+## 交互
+
+1. 右侧面板"快速命令"区域列出所有内置+自定义命令
+2. 每条显示：▶ 图标 + 命令简称（可自定义）+ 命令内容（灰色小字截断）
+3. 点击 → 发送命令到终端（sendCmd 逻辑）
+4. 每条命令右侧 hover 显示编辑/删除按钮
+5. "+ 添加命令" 按钮在最下方
+
+## 数据模型（localStorage）
+
+```
+nexus_terminal_commands = [
+  { id: uuid, name: "查看 Nginx 状态", cmd: "systemctl status nginx" },
+  { id: uuid, name: "读日志", cmd: "tail -f /var/log/nginx/access.log" },
+  ...
+]
+```
+
+## 安全约束
+
+- 纯前端功能，localStorage 存储，不涉及后端
+- 命令发送走已有 WebSocket DATA 通道
+- 转义由终端侧处理，JS 不发 shell 命令（只发字符流）
+
+## 验收标准
+
+1. 右侧面板显示"快速命令"区域
+2. 默认 5 条内置命令可点击执行
+3. 可添加自定义命令（名称+内容）
+4. 可编辑/删除自定义命令
+5. 刷新页面后自定义命令保留
+6. 内置命令不可删除

server/api/websocket.py

@@ -473,13 +473,14 @@ async def broadcast_sync_progress(
     batch_id: str, server_id: int, server_name: str,
     status: str, completed: int, failed: int, total: int,
     error_message: str = None, duration_seconds: int = None,
-    retry_job_id: int = None,
+    retry_job_id: int = None, cancelled: int = 0,
 ):
     """Broadcast per-server push progress to all WebSocket clients.
 
     Called from sync_engine_v2._sync_one() after each server completes.
     Frontend uses batch_id to filter stale events from previous pushes.
     retry_job_id is set when a failed push auto-creates a retry job.
+    cancelled is the count of servers skipped due to user cancellation.
     """
     msg = {
         "type": "sync_progress",
@@ -493,6 +494,7 @@ async def broadcast_sync_progress(
         "error_message": error_message,
         "duration_seconds": duration_seconds,
         "retry_job_id": retry_job_id,
+        "cancelled": cancelled,
     }
     await _dispatch_ws_message(msg)
 

server/application/services/sync_engine_v2.py

@@ -74,12 +74,55 @@ class SyncEngineV2:
         total = len(servers)
         completed = 0
         failed = 0
+        cancelled = 0
         _counter_lock = asyncio.Lock()
         results = {}
 
         async def _sync_one(server: Server):
-            nonlocal completed, failed
+            nonlocal completed, failed, cancelled
             async with sem:
+                # Check cancel flag before starting
+                try:
+                    from server.infrastructure.redis.client import get_redis
+                    redis = get_redis()
+                    cancel_flag = await redis.get(f"sync:cancel:{batch_id}")
+                    if cancel_flag:
+                        sync_log = SyncLog(
+                            server_id=server.id,
+                            source_path=source_path,
+                            target_path=target_path or server.target_path or "/tmp/sync",
+                            trigger_type=trigger_type,
+                            operator=operator,
+                            status="cancelled",
+                            sync_mode=sync_mode,
+                            started_at=datetime.now(timezone.utc),
+                            finished_at=datetime.now(timezone.utc),
+                            error_message="推送已取消",
+                        )
+                        sync_log = await self.sync_log_repo.create(sync_log)
+
+                        async with _counter_lock:
+                            cancelled += 1
+                            try:
+                                from server.api.websocket import broadcast_sync_progress
+                                await broadcast_sync_progress(
+                                    batch_id=batch_id,
+                                    server_id=server.id,
+                                    server_name=server.name,
+                                    status="cancelled",
+                                    completed=completed,
+                                    failed=failed,
+                                    total=total,
+                                    error_message="推送已取消",
+                                )
+                            except Exception as e:
+                                logger.debug(f"WS broadcast failed for cancelled server {server.id}: {e}")
+
+                        results[server.id] = (sync_log, None)
+                        return
+                except Exception as e:
+                    logger.warning(f"Cancel check failed for batch {batch_id}, proceeding with push: {e}")
+
                 dest = target_path or server.target_path or "/tmp/sync"
                 sync_log = SyncLog(
                     server_id=server.id,
@@ -147,6 +190,7 @@ class SyncEngineV2:
                             error_message=sync_log.error_message[:200] if sync_log.error_message else None,
                             duration_seconds=sync_log.duration_seconds,
                             retry_job_id=retry_job_id,
+                            cancelled=cancelled,
                         )
                     except Exception as e:
                         logger.debug(f"WS broadcast failed for server {server.id}: {e}")