App status: errored, but error.log empty.

[extensionsapp]

Expected behaviour

Full error logs.

Actual behaviour

Empty logs.

Software versions used

OS         : Debian 8
node.js    : v5.12.0
PM2        : 1.1.3

PM2 Log output

App fell, but because of what I can not understand, because the error log is empty.

[vmarchaud]

@extensionsapp Could you check the pm2.log file and pastebin it ?

[extensionsapp]

@vmarchaud

2016-07-11 08:00:09: Change detected on path restart.server for app filmi.cc - restarting
2016-07-11 08:00:09: Stopping app:filmi.cc id:15
2016-07-11 08:00:09: App [filmi.cc] with id [15] and pid [16306], exited with code [0] via signal [SIGINT]
2016-07-11 08:00:10: Starting execution sequence in -fork mode- for app name:filmi.cc id:15
2016-07-11 08:00:16: App name:filmi.cc id:15 online
2016-07-11 09:00:10: Change detected on path restart.server for app filmi.cc - restarting
2016-07-11 09:00:10: Stopping app:filmi.cc id:15
2016-07-11 09:00:10: Process with pid 16967 still not killed, retrying...
2016-07-11 09:00:13: Process with pid 16967 still not killed, retrying...
2016-07-11 09:00:13: Process with pid 16967 still alive after 1600ms, sending it SIGKILL now...
2016-07-11 09:00:18: Process with pid 16967 still not killed, retrying...
2016-07-11 09:00:18: Process with pid 16967 could not be killed
2016-07-11 09:00:18: Process with pid 16967 already exists
2016-07-11 09:00:18: Trace: [Error: Process with pid 16967 already exists]

[vmarchaud]

@extensionsapp Apparently you use --watch to restart the server when a file change and the as you can see in the log, there was a problem to kill it, dont know why but if you stop watching for changing file, the issue wont happen again.

[extensionsapp]

@vmarchaud I need the server restarts every hour. Since there is a change in the configuration and should be used.

[vmarchaud]

@extensionsapp Could you try with our development branch (npm v2) npm install Unitech/pm2#development -g && pm2 update

[vmarchaud]

See #2629 we group all issue into one.

[jfoclpf]

I had the same issue and now I realize that the flag watch is true by default. It shouldn't! Now it's working

{
  "apps" : [{
    "name"        : "delp.pt",
    "script"      : "server.js",
    "args"        : "--database",
    "instances"   : "max",
    "watch"       : false,
    "exec_mode"   : "cluster",
    "log_date_format": "DD-MM-YYYY"
  }]
}