zoukankan      html  css  js  c++  java
  • 使用Fabric一键批量部署上线/线上环境监控

    本文讲述如何使用fabric进行批量部署上线的功能

    这个功能对于小应用,可以避免开发部署上线的平台,或者使用linux expect开发不优雅的代码。

    前提条件:

    1、运行fabric脚本的机器和其他机器tcp_port=22端口通

    2、ssh可以登录,你有账号密码

    一、先说批量部署上线

    先上代码,再仔细讲解,脚本如下

    # -*- coding:utf-8 -*-
    from fabric.colors import *
    from fabric.api import *
    from contextlib import contextmanager as _contextmanager
    
    # 自动载入
    env.user='data_monitor'
    env.hosts=['10.93.21.21', '10.93.18.34', '10.93.18.35']
    env.password='datamonitor@123'
    # 手动加入
    env.activate = 'source /home/data_monitor/.bash_profile'
    env.directory = '/home/data_monitor/dmonitor/dmonitor'
    
    
    @_contextmanager
    def virtualenv():
        with cd(env.directory):
            with prefix(env.activate):
                yield
    
    
    @task
    def update():
        with virtualenv():
            run("git pull origin master")
    
    
    @task
    def start():
        with virtualenv():
            run("$(nohup gunicorn --worker-class=gevent dmonitor.wsgi:application -b 0.0.0.0:8009 -w 4 &> /dev/null &) && sleep 1", warn_only=True)
            run("$(nohup python manage.py celery worker -Q high -c 30 &> /dev/null &) && sleep 1 ", warn_only=True)
            run("$(nohup python manage.py celery worker -Q mid -c 30 &> /dev/null &) && sleep 1 ", warn_only=True)
            run("$(nohup python manage.py celery worker -Q low -c 30 &> /dev/null &) && sleep 1", warn_only=True)
    
    
    @task
    def stop():
        with virtualenv():
            run("ps -ef | grep gunicorn | grep -v grep | awk '{print $2}'| xargs kill -9", warn_only=True)
            run("ps -ef | grep celery | grep worker | grep -v grep | awk '{print $2}' | xargs kill -9", warn_only=True)
    
    
    @task
    def deploy():
        update()
        stop()
        start()

    2、线上环境监控

    当然一般线上环境没有用fabric监控的,但是开发环境和测试环境的话,一般都是虚拟机,没有人管你。

    所以自己开发一个小型监控程序,监控一下硬盘cpu内存,或者是一些进程(redis/mysql...),还是挺有用的。

    先上代码

    这个文件是各种task

    import logging
    
    from fabric.api import *
    from fabric.context_managers import *
    from fabric.colors import red, yellow, green
    from common.redis import Redis
    from common.config import redis as redis_config
    
    logger = logging.getLogger(__name__)
    redis = Redis(redis_config.get('ip'), redis_config.get('port'))
    
    
    # hard_disk_monitor, item_name=hard_disk
    @task
    def hard_disk_monitor(item_group, item_name, threshold):
        with settings(hide('warnings', 'running', 'stdout', 'stderr'), parallel=True, warn_only=True):
            host = run('hostname -i')
            hard_disk = run("df -hl | grep /dev/vda3 | awk -F ' ' '{print $5}'")
            print green(host + ':' + hard_disk)
            if int(hard_disk.strip('%')) > threshold:
                redis("lpush %s %s" % (':'.join(['machine', item_group, item_name]), host))
    
    
    # memory_monitor, item_name=memory
    @task
    def memory_monitor(item_group, item_name, threshold):
        with settings(hide('warnings', 'running', 'stdout', 'stderr'), parallel=True, warn_only=True):
            host = run('hostname -i')
            memory = run("cat /proc/meminfo | grep MemFree | awk -F ' ' '{print $2}'")
            print yellow(host + ':' + memory)
            if int(memory.strip()) < threshold:
                redis("lpush %s %s" % (':'.join(['machine', item_group, item_name]), host))
    
    
    # base_services_monitor, item_name != hard_disk or item_name != memory
    @task
    def base_services_monitor(item_group, item_name, threshold):
        with settings(hide('warnings','running','stdout','stderr'),parallel=True,warn_only=True):
            host = run('hostname -i')
            count = run("ps -ef | grep %s | grep -v grep | wc -l" % item_name)
            print red(host + ':' + count)
            if int(count.strip()) != threshold:
                redis("hset %s %s %s" % (':'.join(['machine', item_group, item_name]), host, count))
                redis('incr %s' % ':'.join(['machine', item_group, item_name, host]))
                redis('expire %s 1800' % ':'.join(['machine', item_group, item_name, host]))
    
    
    # restart_services_monitor, item_name = tomcat-7.0.57-mis or item_name = tomcat-httpapi
    @task
    def restart_services_monitor(item_start):
        with settings(hide('warnings', 'running', 'stdout', 'stderr'), parallel=True,warn_only=True):
            host = run('hostname -i')
            run(item_start)
            print green(host + ':' + item_start)

    这个文件是执行task

    # -*- coding:utf-8 -*-
    
    from fabric.api import *
    from fabric.context_managers import *
    execute(monitors.hard_disk_monitor, item_group, item_name, item_threshold,
                            hosts=json.loads(item_param.get('item_hosts')))
                    hosts = self.redis('lrange %s 0 -1' % ':'.join(['machine', item_group, item_name]))
  • 相关阅读:
    最好的我们
    外滩
    外滩
    Java EE (5) -- Java EE 6 JavaServer Faces Developer Certified Expert(1z0-896)
    列表list(序列)、元组tuple(序列)
    P1165 日志分析 洛谷
    T1365 浴火银河星际跳跃 codevs
    T1503 愚蠢的宠物 codevs
    P2820 局域网 洛谷
    T1992 聚会 codevs
  • 原文地址:https://www.cnblogs.com/kangoroo/p/7207782.html
Copyright © 2011-2022 走看看