监控HDFS每天数据增量（Python2实现）

• 查看HDFS数据量
• 数据量监控
• • 1、Python脚本
  • 2、定时任务，每天零点执行
  • 3、查看数据

hadoop fs -df

命令结果截图

命令结果截图（CDH）

touch /root/monitor_hdfs_df.py
chmod 777 /root/monitor_hdfs_df.py
vim /root/monitor_hdfs_df.py

#!/usr/bin/python2
# coding=utf-8
from time import strftime
from os import path
from subprocess import check_output
from re import split
from sys import argv

SEP = ','
SCALING = 70  # 条形图长度缩放规模
FILE_NAME = 'fs-df.txt'  # 存放数据的文件名
file_name = path.join(path.dirname(__file__), FILE_NAME)  # 数据文件的全路径

def fs():
    """查询HDFS数据量"""
    result = check_output('hadoop fs -df', shell=True)
    now = strftime('%Y-%m-%d %H:%M:%S')
    dt = {k: v for k, v in zip(*(split('s+', line) for line in result.strip().split('n')))}
    return now, dt['Used']

def write():
    """把当前HDFS数据量写到文件"""
    with open(file_name, 'a') as f:
        f.write(SEP.join(fs()) + 'n')

def read():
    """读文件"""
    with open(file_name, 'r') as f:
        for line in f.read().strip().split('n'):
            now, used = line.split(SEP)
            yield now, int(used)

def used_bars():
    """查看HDFS已占用的数据量"""
    values = []
    labels = []
    for now, used in read():
        values.append(used)
        labels.append(now)
    # 条形图
    maximum = max(values)
    label_max_length = max(len(label) for label in labels)
    for value, label in zip(values, labels):
        length = int((SCALING * 1.5 * value) / maximum)
        print('{} {} {}'.format('%{}s'.format(label_max_length) % label, '#' * length, value))

def increase_bars():
    """查看HDFS数据增量"""
    lines = list(read())
    values = []
    labels = []
    for (today, used_today), (_, used_yesterday) in zip(lines[1:], lines[:-1]):
        increase = used_today - used_yesterday
        values.append(increase)
        labels.append(today)
    # https://yellow520.blog.csdn.net/article/details/121498190
    get_bar_length = lambda _value, _max_abs: int((abs(_value) * SCALING) / _max_abs)
    labels = labels or [str(i + 1) for i in range(len(values))]
    label_max_length = max(len(label) for label in labels)
    maximum = max(values)
    minimum = min(values)
    max_abs = max(abs(maximum), abs(minimum))
    if max_abs == 0:
        for value, label in zip(values, labels):
            print(label, value)
    else:
        if minimum < 0:
            value_length = len(str(minimum))
            bar_length_left = get_bar_length(minimum, max_abs)
            length_left = value_length + bar_length_left + 1
        else:
            length_left = 0
        for value, label in zip(values, labels):
            mid = ' %{}s '.format(label_max_length) % label
            bar_length = get_bar_length(value, max_abs)
            if value > 0:
                left = ' ' * length_left
                right = '{} {}'.format('#' * bar_length, value)
            elif value < 0:
                left = '{} {}'.format(value, '#' * bar_length).rjust(length_left, ' ')
                right = ''
            else:
                left = ' ' * length_left
                right = '0'
            print(left + mid + right)

def usage():
    print('使用方法：')
    print(__file__ + ' + 数字选项（1、2、3）')
    print('1：' + write.__doc__)
    print('2：' + increase_bars.__doc__)
    print('3：' + used_bars.__doc__)
    exit()

if len(argv) != 2:
    usage()
a = argv[1]
if a == '1':
    write()
elif a == '2':
    increase_bars()
elif a == '3':
    used_bars()
else:
    usage()

添加或删除定时任务（edit）

crontab -e

以vim模式写入并保存

# 每天0点跑一次，记录HDFS数据量
0 0 * * * /root/monitor_hdfs_df.py 1

检查一下定时任务

crontab -l

查看数据量

python monitor_hdfs_df.py 3

查看数据增量

python monitor_hdfs_df.py 2

截图

有负数……