PM2 启动管理多个 app 出现 id 混乱是神马情况？

[tofishes]

首先我使用了不同的启动方式：

pm2 start index.js
pm2 start process.json

启动了 3 个 app ，第一次执行是正常的：

$ pm2 list
┌──────────┬────┬──────┬───────┬────────┬─────────┬────────┬─────────────┬──────────┐
│ App name │ id │ mode │ pid   │ status │ restart │ uptime │ memory      │ watching │
├──────────┼────┼──────┼───────┼────────┼─────────┼────────┼─────────────┼──────────┤
│ index    │ 0  │ fork │ 48337 │ online │ 0       │ 2m     │ 23.746 MB   │ disabled │
│ lianj-pc │ 1  │ fork │ 48871 │ online │ 1       │ 20s    │ 74.781 MB   │ disabled │
│ demo     │ 2  │ fork │ 48889 │ online │ 1       │ 20s    │ 64.785 MB   │ disabled │
└──────────┴────┴──────┴───────┴────────┴─────────┴────────┴─────────────┴──────────┘

其中 index 启动端口为 8001 ，其他两个分别为 9000 ， 8000 。 ** index id=0, lianj-pc id=1, demo id=2 **

那么问题来了，当我再次单独执行 pm2 start process.json 时，启动出现了问题：

# 节选自 pm2.log
2016-07-08 16:10:24: Stopping app:lianj-pc id:1
2016-07-08 16:10:24: App [lianj-pc] with id [1] and pid [48454], exited with code [0] via signal [SIGINT]
2016-07-08 16:10:24: Starting execution sequence in -fork mode- for app name:lianj-pc id:1
2016-07-08 16:10:24: ***App name:lianj-pc id:1 online***
2016-07-08 16:10:24: Stopping app:demo id:2
2016-07-08 16:10:24: App [demo] with id [2] and pid [48456], exited with code [0] via signal [SIGINT]
2016-07-08 16:10:24: Starting execution sequence in -fork mode- for app name:demo id:2
2016-07-08 16:10:24: ***App name:demo id:2 online***
2016-07-08 16:11:04: Stopping app:lianj-pc id:1
2016-07-08 16:11:04: App [lianj-pc] with id [1] and pid [48871], exited with code [0] via signal [SIGINT]
2016-07-08 16:11:05: Starting execution sequence in -fork mode- for app name:lianj-pc id:0
2016-07-08 16:11:05: ***App name:lianj-pc id:0 online***
2016-07-08 16:11:05: Stopping app:demo id:2
2016-07-08 16:11:05: App [demo] with id [2] and pid [48889], exited with code [0] via signal [SIGINT]
2016-07-08 16:11:05: Starting execution sequence in -fork mode- for app name:demo id:0
2016-07-08 16:11:05: ***App name:demo id:0 online***
2016-07-08 16:11:05: App [lianj-pc] with id [0] and pid [49126], exited with code [255] via signal [SIGINT]
2016-07-08 16:11:05: Starting execution sequence in -fork mode- for app name:demo id:0
2016-07-08 16:11:05: App name:demo id:0 online
2016-07-08 16:11:05: App [demo] with id [0] and pid [49144], exited with code [255] via signal [SIGINT]
2016-07-08 16:11:05: Starting execution sequence in -fork mode- for app name:demo id:0
2016-07-08 16:11:05: App name:demo id:0 online
2016-07-08 16:11:05: App [demo] with id [0] and pid [49158], exited with code [255] via signal [SIGINT]
2016-07-08 16:11:05: Starting execution sequence in -fork mode- for app name:demo id:0
2016-07-08 16:11:05: App name:demo id:0 online
2016-07-08 16:11:05: App [demo] with id [0] and pid [49183], exited with code [255] via signal [SIGINT]
2016-07-08 16:11:05: Starting execution sequence in -fork mode- for app name:demo id:0
2016-07-08 16:11:05: App name:demo id:0 online
2016-07-08 16:11:05: App [demo] with id [0] and pid [49188], exited with code [255] via signal [SIGINT]
"pm2.log" 11301L, 938877C
2016-07-08 16:22:21: App [lianj-pc] with id [0] and pid [17471], exited with code [255] via signal [SIGINT]
2016-07-08 16:22:21: Starting execution sequence in -fork mode- for app name:lianj-pc id:0

请看我在日志中加了***的部分，启动 id 和 app name 的对应已经发生混乱，都对应到了 index 的 id 上去，结果就是 index 不断重启，但是又因为重复绑定端口而报错，日志如下：

M2 App [lianj-pc] with id [0] and pid [66856], exited with code [255] via signal [SIGINT]
PM2 Starting execution sequence in -fork mode- for app name:lianj-pc id:0
PM2 App name:lianj-pc id:0 online
lianj-pc-0 Error: listen EADDRINUSE 127.0.0.1:8001
lianj-pc-0     at Object.exports._errnoException (util.js:870:11)
lianj-pc-0     at exports._exceptionWithHostPort (util.js:893:20)
lianj-pc-0     at Server._listen2 (net.js:1234:14)
lianj-pc-0     at listen (net.js:1270:10)
lianj-pc-0     at net.js:1379:9
lianj-pc-0     at GetAddrInfoReqWrap.asyncCallback [as callback] (dns.js:64:16)
lianj-pc-0     at GetAddrInfoReqWrap.onlookup [as oncomplete] (dns.js:83:10)

想从 Google 问个究竟，却始终查不出什么，不知道大家有没有遇到的，目前来看会不会就是不能混合使用启动方式呢？

补充：

pm2 启动的 index 提供了一种网络服务，当访问 index 时，执行一个 shell 命令，该 shell 命令用于执行 另外两个 app 的重启。

多次反复测试后，用 pm2 启动的 app 再执行 shell 命令用于重启 pm2 就发生混乱。