通过crond自动运行Python脚本实现多台linux服务器的监控

一、Pexpect简介

　　Pexpect 是一个用来启动子程序并对其进行自动控制的 Python 模块，它可以用来和像 ssh、ftp、passwd、telnet 等命令行程序进行自动交互。本文介绍 Pexpect 的主要用法和在实际应用中的注意点。 Python 语言的爱好者，系统管理人员，部署及测试人员都能使用 Pexpect 在自己的工作中实现与命令行交互的自动化。

　　具体可以参考http://www.noah.org/wiki/Pexpect#Download_and_Installation

　　下载地址 http://pexpect.sourceforge.net/pexpect-2.3.tar.gz

　　今天介绍的内容就是利用pexpect来实现的远程多服务器的管理。

二、Python脚本

　　以下是python的源代码。程序的大概过程是使用pexpect的ssh命令循环登陆到远程服务器上，将服务和进程的运行情况纪录到临时文件中，然后再从文件中获取服务和进程的运行情况，如果服务或进程异常停止，程序则将其重新启动，所有服务器的运行状况都将纪录在日志文件中。

# coding=utf-8
#!/usr/bin/env python


import pexpect
import getpass, os
import string
import time
from datetime import datetime, date

ssh_newkey = 'Are you sure you want to continue connecting (yes/no)?'
hosts = []

hosts.append('xx.xx.xx.xx')
user = 'user'
password = 'pwd'

logFile = '/tmp/pexpect-2.3/monitor.log'

def restartService(host, user, password, service):
    child = pexpect.spawn('ssh %s@%s' %(user, host))
    i = child.expect([pexpect.TIMEOUT, ssh_newkey, 'password: '])
    if i == 0: # Timeout
        return None
    if i == 1:
        child.sendline('yes')
    child.sendline(password)
    child.sendline('sudo service %s restart' %service)
    j = child.expect(['Password: ', '$', '#'])
    if j == 0:
        child.sendline(password)
    if j == 1:
        child.sendline(password)
    time.sleep(30)
    child.sendline('exit')
    fout = file(logFile, 'a')
    child.logfile_read = fout
    return child
# restart one process
def restartProcess(host, user, password):
    child = pexpect.spawn('ssh %s@%s' %(user, host))
    i = child.expect([pexpect.TIMEOUT, ssh_newkey, 'password: '])
    if i == 0: # Timeout
        return None
    if i == 1:
        child.sendline('yes')
    child.sendline(password)
    child.sendline('sudo su')
    j = child.expect(['Password: ', '$', '#'])
    if j == 0:
        child.sendline(password)
    if j == 1:
        child.sendline(password)
    child.sendline('cd /processlocation/bin')
    child.sendline('./processName.sh &')
    time.sleep(20)
    child.sendline('exit')
    child.sendline('exit')
    fout = file(logFile, 'a')
    child.logfile_read = fout
    return child

# log process' status to file tmpProc.txt
def logProcessInfo(host, user, password, process):
    child = pexpect.spawn('ssh %s@%s' %(user, host))
    i = child.expect([pexpect.TIMEOUT, ssh_newkey, 'password: '])
    if i == 0: # Timeout
        return None
    if i == 1:
        child.sendline('yes')
    child.sendline(password)
    child.sendline('ps -ef|grep %s' %process)
    child.sendline('exit')
    fout = file('tmpProc.txt', 'w')
    child.logfile_read = fout
    return child
# get process' id from file tmpProc.txt
def getProcessId():
    val = ''
    f = file('tmpProc.txt', 'r')
    f.seek(0)
    while True:
        line = f.readline()
        if line.find('processName') > 0:
            index = line.find(' ')
            subline = line[index+1:]
            while True:
                if subline[0] == ' ':
                    index = subline.find(' ')
                    subline = subline[index+1:]
                else:
                    break
            index2 = subline.find(' ')
            number = subline[:index2]
            val = number
        if len(line) == 0:
            break
    f.close()
    return val

# log service's status to file tmpServ.txt
def logServiceInfo(host, user, password, serviceName):
    child = pexpect.spawn('ssh %s@%s' %(user, host))
    i = child.expect([pexpect.TIMEOUT, ssh_newkey, 'password: '])
    if i == 0: # Timeout
        return None
    if i == 1:
        child.sendline('yes')
    child.sendline(password)
    child.sendline('service %s status' %serviceName)
    child.sendline('exit')
    fout = file('tmpServ.txt', 'w')
    child.logfile_read = fout
    return child    

# get servicee's status of start or stop from file tmpServ.txt
def hasServiceStart():
    val = True
    f = file('tmpServ.txt', 'r')
    f.seek(0)
    while True:
        line = f.readline()
        if line.find('stop') > 0:
            val = False
        if len(line) == 0:
            break
    f.close()
    return val
    
def logOperation(content):
    f = file(logFile, 'a')
    f.write(content)
    f.close()

current = datetime.now()
logOperation('\n\n***')
logOperation(current.strftime("%a %b %d %H:%M:%S %Y") + '\n')    
for host in hosts:
    processChild = logProcessInfo(host, user, password, 'processName')
    processChild.expect(pexpect.EOF)
    processId = getProcessId()
    if processId == '':
        logOperation('processName on ' + host + ' has been stopped\n')
        processChild = restartProcess(host, user, password)
        processChild.expect(pexpect.EOF)
    else:
        logOperation('processName on ' + host + ' is ok\n')
    
    serviceChild = logServiceInfo(host, user, password, 'serviceName')
    serviceChild.expect(pexpect.EOF)
    if hasServiceStart():
        logOperation('serviceName on ' + host + ' is ok\n')
    else:
        logOperation('serviceName on ' + host + ' has been stopped\n')
        serviceChild = restartService(host, user, password, 'serviceName')
        serviceChild.expect(pexpect.EOF)

print 'mission accomplished'

 

三、通过配置crond使脚本能够定时执行

1)登陆到脚本所在的linux服务器上

2)运行命令crontab -e

说明:系统默认的编辑器是VIM,如果不是请加上以下shell:

$EDITOR=vi

$export EDITOR

3)添加0 * * * * python /tmp/pexpect-2.3/autoMonitor.py
4)重起crond
cd /etc/init.d
./crond restart
这样每个小时0分的时候105上会自动运行autoMonitor.py脚本

具体crond的配置可以参考http://www.oklinux.cn/html/Basic/jyjq/20090107/66512.html