best way to setup a webhook to restart apache for a django server

I first tried to use django and then django-webhooks to call a shell script that restarts the server. This didn't work, because the webpage hangs when the server restart is called, as django is reloaded.

Then I used fastcgi and python alone to create a URL that calls the shell script. I know the python script works when I run it on the server, but not when it is run from the URL.

Apache is setup as:

<VirtualHost *:80>
    ServerName webhooks.myserver.com

    DocumentRoot /home/ubuntu/web/common/www
    <Directory />
        Options FollowSymLinks +ExecCGI
        AllowOverride All
    </Directory>

    <Files post.py>
        SetHandler fastcgi-script
    </Files>

    FastCgiServer /home/ubuntu/web/common/www/post.py -processes 2 -socket /tmp/fcgi.sock
</VirtualHost>

The python code called by apache is:

#!/usr/bin/python

import fcgi, warnings, os, subprocess

BASE_DIR = os.getcwd()

def app(environ, start_response):
    cmd = "sudo %s/../deploy/postwebhook.sh >> /var/log/votizen/webhooks_run.log 2>> /var/log/votizen/webhooks_error.log &" % BASE_DIR
    warnings.warn("Running cmd=%s" % cmd)
    bufsize = -1
    PIPE = subprocess.PIPE
    subprocess.Popen(cmd, shell=isinstance(cmd, basestring),
                         bufsize=bufsize, stdin=PIPE, stdout=PIPE,
                         stderr=PIPE, close_fds=True)

    warnings.warn("Post deployment webhook completed")

    start_response('200 OK', [('Content-Type', 'text/html')])
    return('Hello World!')

fcgi.WSGIServer(app, bindAddress = '/tmp/fcgi开发者_StackOverflow中文版.sock').run()

And the shell script is:

#!/bin/bash

# restart the apache server
echo ' '
echo 'post webhooks started'
date '+%H:%M:%S %d-%m-%y'
apache2ctl -t; sudo /etc/init.d/apache2 stop; sudo /etc/init.d/apache2 start

# todo: check if apache failed

# copy media files for apps
echo "moving SC to S3"
python /home/ubuntu/web/corporate/manage.py sync_media_s3 -p sc

date '+%H:%M:%S %d-%m-%y'
echo 'post webhooks completed'

I'm not seeing any errors in the apache logs and the access log shows that the triggering URL is being called. However, I only see the python warnings the first time the URL is called after a restart and it never actually restarts the server.

I'm using webfaction, and the following script works for me:

import time
import os

BASE_DIR = "/path/to/your/app/"

def stopit(app):
    os.popen( os.path.join(BASE_DIR, "apache2", "bin", "stop") )

def startit(app):
    os.popen( os.path.join(BASE_DIR, "apache2", "bin", "start") )

def restart(app):
    stopit(app)
    time.sleep(1)
    startit(app)

The "stop" script then looks like this:

#!/usr/local/bin/python
import os
lines = os.popen('ps -u username -o pid,command').readlines()
running = False
for line in lines:
    if '/path/to/app/apache2/conf/httpd.conf' in line:
        running = True
        proc_id = line.split()[0]
        os.system('kill %s 2> /dev/null' % proc_id)
if not running:
    print "Not running"
else:
    print "Stopped"

The "start" script:

/path/to/app/apache2/bin/httpd -f /path/to/app/apache2/conf/httpd.conf

I didn't use django, but with simple python following code works for me.

import os
os.system('apachectl -k restart')