我在webfaction上运行了一个django应用程序,几天前我升级了一个软件包,该软件包也将django升级到了2.1版本(从2.0版本(。此后不久,芹菜停止工作,当我尝试在主管中重新启动它时,出现以下错误:
http://127.0.0.1:12637 refused connection
当我检查主管日志时,我看到以下内容:
2018-10-06 21:56:58,356 WARN received SIGHUP indicating restart request
2018-10-06 21:56:58,417 INFO waiting for accconsole_celery, catdemo, accconsole_celerybeat to die
2018-10-06 21:56:58,418 INFO exited: catdemo (terminated by SIGTERM; not expected)
2018-10-06 21:56:58,432 WARN received SIGTERM indicating exit request
2018-10-06 21:56:58,590 INFO stopped: accconsole_celerybeat (terminated by SIGHUP)
2018-10-06 21:57:01,604 INFO waiting for accconsole_celery to die
2018-10-06 21:57:04,721 INFO waiting for accconsole_celery to die
2018-10-06 21:57:07,729 INFO waiting for accconsole_celery to die
2018-10-06 21:57:10,736 INFO waiting for accconsole_celery to die
2018-10-06 21:57:13,740 INFO waiting for accconsole_celery to die
2018-10-06 21:57:16,747 INFO waiting for accconsole_celery to die
2018-10-06 21:57:19,752 INFO waiting for accconsole_celery to die
2018-10-06 21:57:22,758 INFO waiting for accconsole_celery to die
2018-10-06 21:57:25,762 INFO waiting for accconsole_celery to die
2018-10-06 21:57:28,766 INFO waiting for accconsole_celery to die
2018-10-06 21:57:31,769 INFO waiting for accconsole_celery to die
2018-10-06 21:57:34,774 INFO waiting for accconsole_celery to die
2018-10-06 21:57:37,778 INFO waiting for accconsole_celery to die
2018-10-06 21:57:40,781 INFO waiting for accconsole_celery to die
2018-10-06 21:57:43,784 INFO waiting for accconsole_celery to die
所以我想有些进程没有正确关闭,但我似乎不能强迫它退出。每次我尝试:ps aux|grep 'celery worker'
我得到一个具有不同流程编号的工人,如下所示:
acounsel 11386 0.0 0.0 112704 980 pts/1 S+ 20:48 0:00 grep --color=auto celery worker
acounsel 11446 0.0 0.0 112704 980 pts/1 S+ 20:49 0:00 grep --color=auto celery worker
所以当我试图杀死他们时,我得到这样的回应:-bash: kill: (11446) - No such process
pkill -9 -f 'celery worker'
和更长的ps auxww | grep 'celery worker' | awk '{print $2}' | xargs kill -9
似乎也无济于事。我也无法运行像celery -A django_aconsole status
或celery -A django_aconsole purge
这样的芹菜命令,它只是挂起并且似乎什么也没做。
任何想法如何关闭这些进程以便重新启动芹菜?
这是一个简单的问题; supervisord 已经停止运行,所以 supervisorctl 没有它就无法运行。只需要先运行主管,一切都像雨一样顺利。