node.js - Node 服务器在永远停止后运行

Question

我有一个 Node 应用程序作为守护进程在服务器上永远运行。在其中一个更新之后，我试图停止它(稍后重新启动)。但令我惊讶的是 forever stopall 什么也没做 No forever processes running 被返回。 Forever list 已返回相同的值。我也用 sudo 试过这两个命令。

问题是我可以清楚地看到 Node 仍在运行(我的应用程序运行良好)。知道出了什么问题吗？

P.S.我需要尽快更新，所以我只是重新启动了服务器。但我还是对这种情况很好奇。

P.S.S 后输入 ps aux | grep app.js

ubuntu    1320  0.0  2.2 663040 23232 ?        Ssl  Sep12   0:00 /usr/bin/nodejs /usr/lib/node_modules/forever/bin/monitor node_app/app.js
ubuntu    1322  0.0  6.9 992564 70792 ?        Sl   Sep12   0:31 /usr/bin/nodejs /var/www/node_app/app.js
root      9739  0.0  0.0  10468   936 pts/0    S+   11:09   0:00 grep --color=auto app.js

为什么会这样？我在亚马逊 AWS 上运行 node app.js。

Answer 1

解决方案 1 - 使用 --uid

I´ve managed how to mark your apps in forever using a UID, and then not using forever stopall. This is a cleaner way, and will make forever to kill all the processes depending on the script.

You just need to add the --uid "scriptID" parameter, and then all the processes depending on it, will be controlled together.

To start a new daemon: forever start --uid "demo" --sourceDir /home/procdir -l /home/log/logfile -o /home/log/outputfile -a -d -v taskName

To Stop a daemon: forever stop -uid "demo"

-bash-4.1$ forever list
info:    Forever processes running
data:        uid       command       script           forever pid   id logfile                                     uptime          
data:    [1] Test /usr/bin/node grunt serve:test 18217   18224    /home/admin/logs/test/forever.log 59:20:21:10.512 
data:    [2] Dev  /usr/bin/node grunt serve:dev 18347   18354    /home/admin/logs/dev/forever.log  59:20:19:56.87  
data:    [3] Prod /usr/bin/node grunt serve:prod 20411   20418    /home/admin/logs/prod/forever.log           59:18:58:28.697 

Anyway, you can also run (hands way) this command to kill the processes: First, kill all forever tasks (this will prevent forever to run the task again when killing it):

 forever list | grep your_app | `awk '/\[0\]/{print "forever stop "$8}'` 

After that, when forever is killed, then now it´s time to kill your node_app

ps -efa | grep  node | grep your_app | `awk '{ print "kill "$2}'` 

I strongly recommend you to not using kill. forever --uid will surely be the best solution.

希望此解决方案对您有所帮助!

解决方案 2 - 永远 --uid 已弃用

As forever --uid was deprecated, I´ve tried to find the way of managing multiple apps by name (without using the --uid deprecated method). It seems using config files is the solution. As read in forever docs uid and id parameters are still in use in config files. After some tests, I´ve managed that id is the right parameter:

例子:

-bash-4.1$ pwd
/jome/myuser/app
-bash-4.1$ forever list
info:    No forever processes running
-bash-4.1$ forever start ./forever/development.json 
warn:    --minUptime not set. Defaulting to: 1000ms
warn:    --spinSleepTime not set. Your script will exit if it does not stay up for at least 1000ms
info:    Forever processing file: index.js

-bash-4.1$ forever stop app4U
info:    Forever stopped process:
    uid  command             script   forever pid   id    logfile                          uptime      
[0] BaoO /usr/local/bin/node index.js 41196   41197 app4U /Users/me/.forever/BaoO.log 0:0:0:5.995 

forever/development.json

{
    // Comments are supported
    "id": "app4U",
    "append": true,
    "watch": true,
    "script": "index.js",
    "sourceDir": "/jome/myuser/app"
}

index.js

a=0;
while (a<10) {
    a+=1
    a-=1
}