GatewayWorker分開部署，過一段時間不用就會停止連接，無法執(zhí)行BusinessWorker
您好，請教一個問題：
?
部署情況：
GatewayWorker的Register?Gateway BusinessWorker全部是分開部署，分別啟動的。全部在部署是一臺本機上。
?
問題描述：
我的BusinessWorker過一段時間不執(zhí)行任務會跟Gateway斷開，無法執(zhí)行后臺任務。具體間隔的時長不太清楚，一般是隔一個晚上，第二天執(zhí)行異步任務，就無法執(zhí)行，必須Reload就可以再次執(zhí)行了。只有一直有任務執(zhí)行就不會出現(xiàn)斷開的情況。
?
日志截圖：
==> workerman.log <==
2019-02-15 23:07:25 pid:13913 Workerman stop success
2019-02-15 23:07:26 pid:12583 SendBufferToWorker fail. The connections between Gateway and BusinessWorker are not ready. See http://wiki.workerman.net/Error3
2019-02-15 23:07:26 pid:12583 SendBufferToWorker fail. The connections between Gateway and BusinessWorker are not ready. See http://wiki.workerman.net/Error3
2019-02-15 23:38:32 pid:13913 Workerman stopping ...
2019-02-15 23:38:32 pid:13913 Workerman has been stopped
2019-02-15 23:38:34 pid:13994 Workerman restart
2019-02-15 23:38:34 pid:13994 Workerman is stopping ...
2019-02-15 23:38:34 pid:13994 Workerman stop success
2019-02-16 08:35:34 pid:16411 Workerman reload
2019-02-16 08:35:34 pid:13997 Workerman reloading3個回答
打賞
您好 Walkor
今天早上Businessworker依舊不執(zhí)行（這個問題，我去年就碰到,當時只能用linux定時任務做Reload）
我確定沒有做任何外部接口調用，因為下半夜沒有任何業(yè)務執(zhí)行，感謝您的再次回復！
?
補充一下我用的是系統(tǒng)環(huán)境?CentOS Linux release 7.4.1708 (Core)? ?64位
Status信息如下：（以下信息是沒有做Reload之前的信息截圖）Workerman status
----------------------------------------------GLOBAL STATUS----------------------------------------------------
Workerman version:3.5.14 PHP version:7.2.12
start time:2019-02-16 22:09:03 run 0 days 10 hours
load average: 8.87, 9, 9 event-loop:\Workerman\Events\Event
2 workers 11 processes
worker_name exit_status exit_count
FileMonitor 0 0
commonBusinessworker 0 0
----------------------------------------------PROCESS STATUS---------------------------------------------------
pid memory listening worker_name connections send_fail timers total_request qps status
10845 2M none FileMonitor 0 0 0 0 0
10846 4M none commonBusinessworker 5 0 0 10 0
10847 4M none commonBusinessworker 5 0 0 4 0
10848 4M none commonBusinessworker 5 0 0 1 0
10849 4M none commonBusinessworker 5 0 0 4 0
10850 4M none commonBusinessworker 5 0 0 101 0
10851 4M none commonBusinessworker 5 0 0 6 0
10852 4M none commonBusinessworker 5 0 0 16 0
10853 4M none commonBusinessworker 5 0 0 1 0
10854 4M none commonBusinessworker 5 0 0 7 0
10855 4M none commonBusinessworker 5 0 0 1 0
----------------------------------------------PROCESS STATUS---------------------------------------------------
Summary 42M - - 50 0 0 151 0 上面Status截圖之后，我Reload一下,異步任務就立刻有效執(zhí)行，日志詳情如下：
2019-02-16 20:34:50 pid:10680 Workerman stopping ...
2019-02-16 20:34:50 pid:10680 Workerman has been stopped
2019-02-16 20:34:50 pid:10692 Workerman stop success
2019-02-16 20:35:42 pid:10707 Workerman status
2019-02-16 20:36:00 pid:10708 Workerman status
2019-02-16 20:36:23 pid:10710 Workerman status
2019-02-16 21:18:33 pid:10771 Workerman status
2019-02-16 21:19:25 pid:10774 Workerman status
2019-02-16 22:09:03 pid:10842 Workerman restart
2019-02-16 22:09:03 pid:10842 Workerman is stopping ...
2019-02-16 22:09:03 pid:10694 Workerman stopping ...
2019-02-16 22:09:03 pid:10694 Workerman has been stopped
2019-02-16 22:09:03 pid:10842 Workerman stop success
2019-02-16 22:09:19 pid:10856 Workerman status
2019-02-17 08:18:32 pid:13297 Workerman status
2019-02-17 08:26:18 pid:13451 Workerman reload
2019-02-17 08:26:18 pid:10844 Workerman reloading
- 暫無評論
從status里看BusinessWorker看是開了4個Gateway進程，如果是4個Gateway進程那么BusinessWorker看起來status是正常的。但是服務器的負載看起來有點高，top看下是什么進程造成這么高的負載。
?還有分部署部署的示意圖貼下，哪些ip放了哪些服務，每個服務的配置貼下。另外描述下具體的業(yè)務是什么？
?
BusinessWorker過一段時間不執(zhí)行任務會跟Gateway斷開
怎么知道的BusinessWorker與Gateway 斷開了？
?
無法執(zhí)行后臺任務
什么異步任務？如何確定的無法執(zhí)行后臺任務？還有能否貼下Events.php代碼？
?
?
-
houniao506 2019-03-28
經過一段時間的排查和觀察，之所以不執(zhí)行任務，是出在Redis和Mysql的自動斷開的原因，他們的引用寫在onWorkerStart里。后來改掉了就好了
感謝大大的及時回復！
