fix: remote agent exec errors
This commit is contained in:
+79
-13
@@ -101,17 +101,21 @@ type Agent struct {
|
|||||||
wsConn *websocket.Conn
|
wsConn *websocket.Conn
|
||||||
wsMu sync.Mutex
|
wsMu sync.Mutex
|
||||||
stopCh chan struct{}
|
stopCh chan struct{}
|
||||||
wsStopCh chan struct{} // 用于停止当前 WebSocket 相关的 goroutine
|
wsStopCh chan struct{} // 用于停止当前 WebSocket 相关的 goroutine
|
||||||
|
taskLogs map[uint][]string // 记录最近的日志行,用于失败显示
|
||||||
|
logMu sync.Mutex // taskLogs 的锁
|
||||||
}
|
}
|
||||||
|
|
||||||
func NewAgent(config *Config, configFile string) *Agent {
|
func NewAgent(config *Config, configFile string) *Agent {
|
||||||
a := &Agent{
|
a := &Agent{
|
||||||
config: config,
|
config: config,
|
||||||
configFile: configFile,
|
configFile: configFile,
|
||||||
machineID: utils.GenerateMachineID(),
|
machineID: utils.GenerateMachineID(),
|
||||||
tasks: make(map[uint]*AgentTask),
|
tasks: make(map[uint]*AgentTask),
|
||||||
client: &http.Client{Timeout: 30 * time.Second},
|
client: &http.Client{Timeout: 30 * time.Second},
|
||||||
stopCh: make(chan struct{}),
|
stopCh: make(chan struct{}),
|
||||||
|
lastTaskCount: -1,
|
||||||
|
taskLogs: make(map[uint][]string),
|
||||||
}
|
}
|
||||||
|
|
||||||
// 初始化调度器
|
// 初始化调度器
|
||||||
@@ -152,6 +156,12 @@ func (h *AgentHandler) OnTaskHeartbeat(req *executor.ExecutionRequest, duration
|
|||||||
"duration": duration,
|
"duration": duration,
|
||||||
})
|
})
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// 每分钟打印一次任务还在运行的日志,提升长任务的存在感
|
||||||
|
if duration >= 60000 && (duration/60000 > (duration-3000)/60000) {
|
||||||
|
logger.Infof("[Scheduler] 任务 #%s 仍在运行中... (已耗时: %v)",
|
||||||
|
req.TaskID, (time.Duration(duration) * time.Millisecond).Round(time.Second))
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
func (h *AgentHandler) OnTaskStarted(req *executor.ExecutionRequest) {}
|
func (h *AgentHandler) OnTaskStarted(req *executor.ExecutionRequest) {}
|
||||||
@@ -171,6 +181,11 @@ func (h *AgentHandler) OnTaskCompleted(req *executor.ExecutionRequest, result *e
|
|||||||
StartTime: result.StartTime.Unix(),
|
StartTime: result.StartTime.Unix(),
|
||||||
EndTime: result.EndTime.Unix(),
|
EndTime: result.EndTime.Unix(),
|
||||||
})
|
})
|
||||||
|
|
||||||
|
if result.Status == "failed" {
|
||||||
|
h.agent.printLastLogs(result.LogID)
|
||||||
|
}
|
||||||
|
h.agent.clearTaskLog(result.LogID)
|
||||||
}
|
}
|
||||||
|
|
||||||
func (h *AgentHandler) OnTaskFailed(req *executor.ExecutionRequest, err error) {
|
func (h *AgentHandler) OnTaskFailed(req *executor.ExecutionRequest, err error) {
|
||||||
@@ -195,6 +210,9 @@ func (h *AgentHandler) OnTaskFailed(req *executor.ExecutionRequest, err error) {
|
|||||||
StartTime: time.Now().Unix(),
|
StartTime: time.Now().Unix(),
|
||||||
EndTime: time.Now().Unix(),
|
EndTime: time.Now().Unix(),
|
||||||
})
|
})
|
||||||
|
|
||||||
|
h.agent.printLastLogs(req.LogID)
|
||||||
|
h.agent.clearTaskLog(req.LogID)
|
||||||
}
|
}
|
||||||
|
|
||||||
func (h *AgentHandler) OnCronNextRun(req *executor.ExecutionRequest, nextRun time.Time) {}
|
func (h *AgentHandler) OnCronNextRun(req *executor.ExecutionRequest, nextRun time.Time) {}
|
||||||
@@ -345,6 +363,7 @@ func (a *Agent) handleWSMessage(msg *WSMessage) {
|
|||||||
}
|
}
|
||||||
|
|
||||||
func (a *Agent) fetchTasks() {
|
func (a *Agent) fetchTasks() {
|
||||||
|
logger.Info("正在从服务器拉取任务列表...")
|
||||||
if err := a.sendWSMessage(WSTypeFetchTasks, map[string]interface{}{}); err != nil {
|
if err := a.sendWSMessage(WSTypeFetchTasks, map[string]interface{}{}); err != nil {
|
||||||
logger.Warnf("请求任务列表失败: %v", err)
|
logger.Warnf("请求任务列表失败: %v", err)
|
||||||
}
|
}
|
||||||
@@ -431,8 +450,8 @@ func (a *Agent) handleTasks(data json.RawMessage) {
|
|||||||
json.Unmarshal(data, &resp)
|
json.Unmarshal(data, &resp)
|
||||||
|
|
||||||
newCount := len(resp.Tasks)
|
newCount := len(resp.Tasks)
|
||||||
if newCount != a.lastTaskCount {
|
if newCount != a.lastTaskCount || newCount == 0 {
|
||||||
logger.Infof("任务列表更新: %d -> %d 个任务", a.lastTaskCount, newCount)
|
logger.Infof("任务列表同步成功: 共获取到 %d 个任务", newCount)
|
||||||
a.lastTaskCount = newCount
|
a.lastTaskCount = newCount
|
||||||
}
|
}
|
||||||
|
|
||||||
@@ -486,6 +505,9 @@ func (w *RealTimeLogWriter) Write(p []byte) (n int, err error) {
|
|||||||
return 0, nil
|
return 0, nil
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// 记录到本地缓存,用于失败时显示
|
||||||
|
w.agent.addTaskLog(w.logID, p)
|
||||||
|
|
||||||
// 构造消息
|
// 构造消息
|
||||||
msg := map[string]interface{}{
|
msg := map[string]interface{}{
|
||||||
"log_id": w.logID,
|
"log_id": w.logID,
|
||||||
@@ -596,7 +618,7 @@ func (a *Agent) updateTasks(tasks []AgentTask) {
|
|||||||
if _, exists := newTasks[id]; !exists {
|
if _, exists := newTasks[id]; !exists {
|
||||||
a.cronManager.RemoveTask(fmt.Sprintf("%d", id))
|
a.cronManager.RemoveTask(fmt.Sprintf("%d", id))
|
||||||
delete(a.tasks, id)
|
delete(a.tasks, id)
|
||||||
logger.Infof("移除任务 #%d", id)
|
logger.Infof("移除调度任务 #%d", id)
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
@@ -607,13 +629,13 @@ func (a *Agent) updateTasks(tasks []AgentTask) {
|
|||||||
if task.Enabled {
|
if task.Enabled {
|
||||||
err := a.cronManager.AddTask(task)
|
err := a.cronManager.AddTask(task)
|
||||||
if err != nil {
|
if err != nil {
|
||||||
logger.Errorf("调度任务 #%d 失败: %v", id, err)
|
logger.Errorf("添加调度任务 #%d 失败: %v", id, err)
|
||||||
continue
|
continue
|
||||||
}
|
}
|
||||||
logger.Infof("已调度任务 #%d %s (%s)", id, task.Name, task.GetSchedule())
|
logger.Infof("已添加调度任务 #%d %s (%s)", id, task.Name, task.GetSchedule())
|
||||||
} else {
|
} else {
|
||||||
a.cronManager.RemoveTask(fmt.Sprintf("%d", id))
|
a.cronManager.RemoveTask(fmt.Sprintf("%d", id))
|
||||||
logger.Infof("任务 #%d 已禁用", id)
|
logger.Infof("调度任务 #%d 已禁用", id)
|
||||||
}
|
}
|
||||||
a.tasks[id] = task
|
a.tasks[id] = task
|
||||||
}
|
}
|
||||||
@@ -634,6 +656,50 @@ func (a *Agent) clearAllTasks() {
|
|||||||
logger.Info("所有任务已清空")
|
logger.Info("所有任务已清空")
|
||||||
}
|
}
|
||||||
|
|
||||||
|
func (a *Agent) addTaskLog(logID uint, p []byte) {
|
||||||
|
if logID == 0 {
|
||||||
|
return
|
||||||
|
}
|
||||||
|
a.logMu.Lock()
|
||||||
|
defer a.logMu.Unlock()
|
||||||
|
|
||||||
|
content := string(p)
|
||||||
|
lines := strings.Split(strings.TrimSuffix(content, "\n"), "\n")
|
||||||
|
|
||||||
|
a.taskLogs[logID] = append(a.taskLogs[logID], lines...)
|
||||||
|
if len(a.taskLogs[logID]) > 50 {
|
||||||
|
a.taskLogs[logID] = a.taskLogs[logID][len(a.taskLogs[logID])-50:]
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
func (a *Agent) printLastLogs(logID uint) {
|
||||||
|
if logID == 0 {
|
||||||
|
return
|
||||||
|
}
|
||||||
|
a.logMu.Lock()
|
||||||
|
lines, ok := a.taskLogs[logID]
|
||||||
|
a.logMu.Unlock()
|
||||||
|
|
||||||
|
if !ok || len(lines) == 0 {
|
||||||
|
return
|
||||||
|
}
|
||||||
|
|
||||||
|
logger.Errorf("--- 任务 #%d 失败日志预览 (最近 %d 行) ---", logID, len(lines))
|
||||||
|
for _, line := range lines {
|
||||||
|
fmt.Println(" " + line)
|
||||||
|
}
|
||||||
|
logger.Errorf("--- 任务 #%d 结束 ---", logID)
|
||||||
|
}
|
||||||
|
|
||||||
|
func (a *Agent) clearTaskLog(logID uint) {
|
||||||
|
if logID == 0 {
|
||||||
|
return
|
||||||
|
}
|
||||||
|
a.logMu.Lock()
|
||||||
|
defer a.logMu.Unlock()
|
||||||
|
delete(a.taskLogs, logID)
|
||||||
|
}
|
||||||
|
|
||||||
// executeTask 已被 AgentHandler.OnTaskCompleted 代替,此处删除旧实现
|
// executeTask 已被 AgentHandler.OnTaskCompleted 代替,此处删除旧实现
|
||||||
|
|
||||||
func (a *Agent) doRequest(method, path string, body interface{}) (*http.Response, error) {
|
func (a *Agent) doRequest(method, path string, body interface{}) (*http.Response, error) {
|
||||||
|
|||||||
@@ -439,6 +439,9 @@ func (c *AgentController) WSConnect(ctx *gin.Context) {
|
|||||||
// 启动读写协程
|
// 启动读写协程
|
||||||
go c.wsWritePump(ac)
|
go c.wsWritePump(ac)
|
||||||
go c.wsReadPump(ac, agent)
|
go c.wsReadPump(ac, agent)
|
||||||
|
|
||||||
|
// 主动推送任务列表
|
||||||
|
go c.wsManager.BroadcastTasks(agent.ID)
|
||||||
}
|
}
|
||||||
|
|
||||||
// wsReadPump 读取消息
|
// wsReadPump 读取消息
|
||||||
|
|||||||
@@ -77,6 +77,11 @@ func (m *CronManager) AddTask(task CronTask) error {
|
|||||||
timeout := task.GetTimeout()
|
timeout := task.GetTimeout()
|
||||||
|
|
||||||
entryID, err := m.cron.AddFunc(task.GetSchedule(), func() {
|
entryID, err := m.cron.AddFunc(task.GetSchedule(), func() {
|
||||||
|
defer func() {
|
||||||
|
if r := recover(); r != nil {
|
||||||
|
m.logger.Errorf("[CronManager] 任务 #%s 执行过程中发生 Panic: %v", taskID, r)
|
||||||
|
}
|
||||||
|
}()
|
||||||
m.logger.Infof("[CronManager] 触发计划任务 #%s (%s)", taskID, name)
|
m.logger.Infof("[CronManager] 触发计划任务 #%s (%s)", taskID, name)
|
||||||
|
|
||||||
req := &ExecutionRequest{
|
req := &ExecutionRequest{
|
||||||
|
|||||||
@@ -262,15 +262,27 @@ func (s *Scheduler) worker(id int) {
|
|||||||
case <-s.stopCh:
|
case <-s.stopCh:
|
||||||
return
|
return
|
||||||
case req := <-s.taskQueue:
|
case req := <-s.taskQueue:
|
||||||
// 速率限制
|
func() {
|
||||||
<-s.rateLimiter
|
defer func() {
|
||||||
s.executeTask(req)
|
if r := recover(); r != nil {
|
||||||
|
s.logger.Errorf("[Scheduler] Worker %d panic while processing task %s: %v", id, req.TaskID, r)
|
||||||
|
}
|
||||||
|
}()
|
||||||
|
// 速率限制
|
||||||
|
<-s.rateLimiter
|
||||||
|
s.executeTask(req)
|
||||||
|
}()
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
// executeTask 执行任务(本地执行)
|
// executeTask 执行任务(本地执行)
|
||||||
func (s *Scheduler) executeTask(req *ExecutionRequest) (*ExecutionResult, error) {
|
func (s *Scheduler) executeTask(req *ExecutionRequest) (*ExecutionResult, error) {
|
||||||
|
defer func() {
|
||||||
|
if r := recover(); r != nil {
|
||||||
|
s.logger.Errorf("[Scheduler] 任务 %s 执行过程中发生 Panic: %v", req.TaskID, r)
|
||||||
|
}
|
||||||
|
}()
|
||||||
start := time.Now()
|
start := time.Now()
|
||||||
|
|
||||||
s.logger.Infof("[Scheduler] 执行任务 %s (名称: %s, 类型: %s)", req.TaskID, req.Name, req.Type)
|
s.logger.Infof("[Scheduler] 执行任务 %s (名称: %s, 类型: %s)", req.TaskID, req.Name, req.Type)
|
||||||
|
|||||||
@@ -271,48 +271,55 @@ func (m *AgentWSManager) cleanupLoop() {
|
|||||||
NewAgentService().ResetAllAgentsToOffline()
|
NewAgentService().ResetAllAgentsToOffline()
|
||||||
|
|
||||||
for range ticker.C {
|
for range ticker.C {
|
||||||
m.mu.Lock()
|
func() {
|
||||||
now := time.Now()
|
defer func() {
|
||||||
|
if r := recover(); r != nil {
|
||||||
|
logger.Errorf("[AgentWS] cleanupLoop panic: %v", r)
|
||||||
|
}
|
||||||
|
}()
|
||||||
|
m.mu.Lock()
|
||||||
|
now := time.Now()
|
||||||
|
|
||||||
// 清理超时连接
|
// 清理超时连接
|
||||||
for agentID, conn := range m.connections {
|
for agentID, conn := range m.connections {
|
||||||
if now.Sub(conn.LastPing) > 2*time.Minute {
|
if now.Sub(conn.LastPing) > 2*time.Minute {
|
||||||
// 减少 IP 连接计数
|
// 减少 IP 连接计数
|
||||||
if conn.IP != "" {
|
if conn.IP != "" {
|
||||||
if count, ok := m.ipConnections[conn.IP]; ok && count > 0 {
|
if count, ok := m.ipConnections[conn.IP]; ok && count > 0 {
|
||||||
m.ipConnections[conn.IP] = count - 1
|
m.ipConnections[conn.IP] = count - 1
|
||||||
|
}
|
||||||
|
}
|
||||||
|
conn.Close()
|
||||||
|
delete(m.connections, agentID)
|
||||||
|
// 更新数据库状态
|
||||||
|
database.DB.Model(&models.Agent{}).Where("id = ?", agentID).Update("status", "offline")
|
||||||
|
logger.Infof("[AgentWS] Agent #%d 心跳超时,已断开", agentID)
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
// 定期清理数据库中的过期状态(处理服务重启或异常终止的情况)
|
||||||
|
// 有些 Agent 虽然没有连接,但数据库状态可能是 "online"
|
||||||
|
cutoff := now.Add(-2 * time.Minute)
|
||||||
|
database.DB.Model(&models.Agent{}).
|
||||||
|
Where("status = ? AND last_seen < ?", "online", cutoff).
|
||||||
|
Update("status", "offline")
|
||||||
|
|
||||||
|
// 清理过期的限流记录(超过 10 分钟未活动)
|
||||||
|
|
||||||
|
// 清理过期的限流记录(超过 10 分钟未活动)
|
||||||
|
for ip, lastAttempt := range m.ipLastAttempt {
|
||||||
|
if now.Sub(lastAttempt) > 10*time.Minute {
|
||||||
|
delete(m.ipLastAttempt, ip)
|
||||||
|
delete(m.ipFailCount, ip)
|
||||||
|
// 只清理没有活跃连接的 IP 计数
|
||||||
|
if m.ipConnections[ip] == 0 {
|
||||||
|
delete(m.ipConnections, ip)
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
conn.Close()
|
|
||||||
delete(m.connections, agentID)
|
|
||||||
// 更新数据库状态
|
|
||||||
database.DB.Model(&models.Agent{}).Where("id = ?", agentID).Update("status", "offline")
|
|
||||||
logger.Infof("[AgentWS] Agent #%d 心跳超时,已断开", agentID)
|
|
||||||
}
|
}
|
||||||
}
|
|
||||||
|
|
||||||
// 定期清理数据库中的过期状态(处理服务重启或异常终止的情况)
|
m.mu.Unlock()
|
||||||
// 有些 Agent 虽然没有连接,但数据库状态可能是 "online"
|
}()
|
||||||
cutoff := now.Add(-2 * time.Minute)
|
|
||||||
database.DB.Model(&models.Agent{}).
|
|
||||||
Where("status = ? AND last_seen < ?", "online", cutoff).
|
|
||||||
Update("status", "offline")
|
|
||||||
|
|
||||||
// 清理过期的限流记录(超过 10 分钟未活动)
|
|
||||||
|
|
||||||
// 清理过期的限流记录(超过 10 分钟未活动)
|
|
||||||
for ip, lastAttempt := range m.ipLastAttempt {
|
|
||||||
if now.Sub(lastAttempt) > 10*time.Minute {
|
|
||||||
delete(m.ipLastAttempt, ip)
|
|
||||||
delete(m.ipFailCount, ip)
|
|
||||||
// 只清理没有活跃连接的 IP 计数
|
|
||||||
if m.ipConnections[ip] == 0 {
|
|
||||||
delete(m.ipConnections, ip)
|
|
||||||
}
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
m.mu.Unlock()
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|||||||
@@ -161,6 +161,12 @@ func (h *ServerSchedulerHandler) OnTaskHeartbeat(req *executor.ExecutionRequest,
|
|||||||
if req.LogID > 0 {
|
if req.LogID > 0 {
|
||||||
h.es.taskLogService.UpdateTaskDuration(req.LogID, duration)
|
h.es.taskLogService.UpdateTaskDuration(req.LogID, duration)
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// 每分钟打印一次任务还在运行的日志
|
||||||
|
if duration >= 60000 && (duration/60000 > (duration-3000)/60000) {
|
||||||
|
logger.Infof("[Scheduler] 任务 #%s 仍在运行中... (已耗时: %v)",
|
||||||
|
req.TaskID, (time.Duration(duration) * time.Millisecond).Round(time.Second))
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
func (h *ServerSchedulerHandler) OnTaskStarted(req *executor.ExecutionRequest) {
|
func (h *ServerSchedulerHandler) OnTaskStarted(req *executor.ExecutionRequest) {
|
||||||
|
|||||||
Reference in New Issue
Block a user