mongodb备份策略

概述

数据库的备份非常非常非常重要！！！否则出问题连哭的机会有没有（欲哭无泪）今天主要是做一个mongodb的数据库备份。

1、关于备份

备份其实很简单，这里选择的是对mongodb中的某个库进行全备，例如saturn。关于备份策略，要根据实际情况来定，我们是一个小时备份一次。不废话直接上代码：

#!/usr/bin/env python

#-*- encoding: utf-8 -*-

import time
from subprocess import call

go_time = time.strftime("%Y-%m-%d_%H")

dir_path = "/data/backup"

def db_backup(user,pwd,host,db):
   
    call("mongodump -u {} -p {} -h {} -d {} --gzip --archive={}/saturn-{}.gz".format(user,pwd,host,db,dir_path,go_time))

if __name__ == '__main__':
   
    db_backup("hqms","hqms123","127.0.0.1:27017","saturn")

备份的文件saturn_2018-01-17_23.gz

我把备份和删除旧文件分开了，why？写在一个里面不是很好吗，主要是两个任务的执行频率不一样，删除的任务执行频率比较低，每天执行一次既可以，否则会浪费系统资源。

并且过多的耦合总会出问题。下面贴出删除代码：

#!/usr/bin/env python
# -*-encoding: utf-8 -*-

import  os
import time

cur_time = time.time()
#当天零点的时间戳
st = cur_time - cur_time%86400 + time.timezone

dir_path = "/data/backup"

#删除
def drop_file(file_dir):
    for filename in os.listdir(file_dir):
        if os.path.isfile(file_dir + "/" +filename):
            ft = os.stat(file_dir + "/" +filename)
            lt = int(ft.st_mtime)
            nt = cur_time - 3600*2  #距离现在2个小时之前的时间戳
            dt = st - 86400*5       #距离现在5天之前的时间戳
           
            #delete  file  by hour in today
            if st < lt < nt:
                try:
                    os.remove(file_dir + "/" + filename)
                    time.sleep(3)
                    print "Deleted success" + file_dir + "/" + filename
                except (SyntaxError,IOError) as error:
                    print error
            #delete file by day
            elif lt < dt:
                try:
                    os.remove(file_dir + "/" + filename)
                    time.sleep(3)
                    print "Deleted success" + file_dir + "/" + filename
                except (SyntaxError,IOError) as error:
                    print error
if __name__ == '__main__':
  
    drop_file(dir_path)

说明删除的代码是python2.7版本的。

附：

python3.6版本：

#!/usr/bin/env python

import  os
import time

cur_time = time.time()
#获取当天零点的时间戳
st = cur_time - cur_time%86400 + time.timezone

dir_path = "/data/backup"

#删除
def drop_file(file_dir):
    for filename in os.listdir(file_dir):
        if os.path.isfile(file_dir + "/" +filename):
            ft = os.stat(file_dir + "/" +filename)
            lt = int(ft.st_mtime)
            nt = cur_time - 3600*2 #获取当天保留截止的时间戳
            dt = st - 86400*5      #获取保留天数的时间戳
           
            #delete  file  by hour in today
            if st < lt < nt:
                try:
                    os.remove(file_dir + "/" + filename)
                    time.sleep(3)
                    print ("Deleted success" + file_dir + "/" + filename)
                except (SyntaxError,IOError) as e:
                    print (e)
            #delete file by day
            elif lt < dt:
                try:
                    os.remove(file_dir + "/" + filename)
                    time.sleep(3)
                    print ("Deleted success" + file_dir + "/" + filename)
                except (SyntaxError,IOError) as e:
                    print (e)
if __name__ == '__main__':
  
    drop_file(dir_path)

2、关于策略

备份脚本是一个小时执行一次，删除脚本是每天执行一次。使用jenkins的pipeline执行，给出删除策略的pipeline：

1）create a pipeline task in jenkins

pipeline{
   agent{ label 'db1'}
   stages{
      stage('drop old file'){
         steps{
    dir('/data/scripts'){
      sh 'python rm_backup.py'
    }
  }
      }
   }
}

2）you must create a schedule
 
 Build periodically
 50  23  *  *  *