dotcloud supervisord twistd 提前退出

Question

我正在尝试在 dotcloud 上设置一个 twistd 守护进程:

我的 supervisord.conf 文件:

[program:apnsd]
command=/home/dotcloud/env/bin/twistd --logfile /var/log/supervisor/apnsd.log apnsd -c gp_config.py
directory=/home/dotcloud/current/apnsd

但是，它看起来像命令“提前退出”，然后提示 Supervisor 尝试重新启动，然后失败，因为 twistd dameon 正在后台运行。

来自监督日志:

more supervisord.log
2012-05-19 03:07:52,723 CRIT Set uid to user 1000
2012-05-19 03:07:52,723 WARN Included extra file "/etc/supervisor/conf.d/uwsgi.c
onf" during parsing
2012-05-19 03:07:52,723 WARN Included extra file "/home/dotcloud/current/supervi
sord.conf" during parsing
2012-05-19 03:07:52,922 INFO RPC interface 'supervisor' initialized
2012-05-19 03:07:52,922 WARN cElementTree not installed, using slower XML parser
 for XML-RPC
2012-05-19 03:07:52,923 CRIT Server 'unix_http_server' running without any HTTP
authentication checking
2012-05-19 03:07:52,932 INFO daemonizing the supervisord process
2012-05-19 03:07:52,934 INFO supervisord started with pid 144
2012-05-19 03:07:53,941 INFO spawned: 'apnsd' with pid 147
2012-05-19 03:07:53,949 INFO spawned: 'uwsgi' with pid 149
2012-05-19 03:07:54,706 INFO exited: apnsd (exit status 0; not expected)
2012-05-19 03:07:55,712 INFO spawned: 'apnsd' with pid 175
2012-05-19 03:07:55,712 INFO success: uwsgi entered RUNNING state, process has s
tayed up for > than 1 seconds (startsecs)
2012-05-19 03:07:56,261 INFO exited: apnsd (exit status 1; not expected)
2012-05-19 03:07:58,267 INFO spawned: 'apnsd' with pid 176
2012-05-19 03:07:58,783 INFO exited: apnsd (exit status 1; not expected)
2012-05-19 03:08:01,790 INFO spawned: 'apnsd' with pid 177
2012-05-19 03:08:02,840 INFO success: apnsd entered RUNNING state, process has s
tayed up for > than 1 seconds (startsecs)

来自 apnsd 日志:

dotcloud@hack-default-www-0:/var/log/supervisor$ more apnsd-stderr---supervisor
-m7GnKV.log
INFO:root:Reactor Type: <twisted.internet.pollreactor.PollReactor object at 0x10
a09d0>
DEBUG:root:Creating listener: apnsd.listeners.line.LineProtocolFactory
INFO:root:Listening on Line Protocol on :1055
DEBUG:root:Listener Created: <apnsd.listeners.line.LineProtocolFactory instance
at 0x12fc8c0>
DEBUG:root:Creating App Factory: apnsd.daemon.APNSFactory
INFO:root:Connecting to APNS Server, App: apns_dev:AAA.com.company.www
INFO:root:apns_dev:AAA.com.company.www -> Started connecting to APNS con
nector...
INFO:root:Registering Application: apns_dev:GoParcel...
DEBUG:root:Creating App Factory: apnsd.daemon.APNSFactory
INFO:root:Connecting to APNS Server, App: apns_dev:T365ED94A9.com.appitems.parce
ls
INFO:root:apns_dev:T365ED94A9.com.appitems.parcels -> Started connecting to APNS
 connector...
INFO:root:Registering Application: apns_dev:GoParcelVictor...
Another twistd server is running, PID 172

This could either be a previously started instance of your application or a
different application entirely. To start a new one, either run it in some other
directory, or use the --pidfile and --logfile parameters to avoid clashes.

Another twistd server is running, PID 172
--More--(42%)

worker的状态是failed:

./dotcloud run hack.worker supervisorctl status
USER PATH IS:  C:\Users\Taras/.dotcloud\dotcloud.conf
# supervisorctl status
apnsd                            FATAL      Exited too quickly (process log may
have details)

但是 twistd 进程在那里(ps -ef):

dotcloud   171     1  0 03:13 ?        00:00:00 /home/dotcloud/env/bin/python /home/dotcloud/env/bin/twistd --logfile /var/log/supervisor/apnsd.log apnsd -c gp_config.py

我在尝试通过包装脚本启动进程时遇到了类似的问题(并使用 exec 以便不创建子进程)。我做错了什么？

Answer 1

Supervisor 希望受控进程保留在前台，但 twistd 默认 fork 到后台。 Supervisor因此认为它已经退出，并尝试重新启动它。

您应该使用--nodaemon 选项启动twistd:twistd 将保留在前台，Supervisor 会很高兴!