iops概念主要用在数据方面，这个指标是数据库性能评定的一个重要参考，iops的是每秒进行读写（I/O）操作的次数，主要看随机访问的性能，一般为了iops增高都要依靠磁盘阵列，实际线上的数据库基本都是raid10的配置，raid5在实际生产环境中如果压力上来是抗不住的，当然也要开具体业务压力情况，如果是用物理机就要看iops在实际中能跑到多少值，现在云也普遍了，如果你用的RDS云数据库，这个iops是可以根据业务情况自己选择的，基本是个参数，可以按需进行修改，当然数值越大费用越高，今天主要是用python通过系统数据来获取磁盘的iops，便于监控使用情况，代码如下：
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
#!/usr/bin/python import os, time, math run_tests = 3 devices = os.listdir('/sys/block/') check_devices = [] reads = {} writes = {} for dev in devices: if dev.startswith('md') or dev.startswith('sd') or dev.startswith('hd'): check_devices.append(dev) reads[dev] = [] writes[dev] = [] check_devices = sorted(check_devices) for t in range(run_tests + 1): for dev in check_devices: file_data = open('/sys/block/%s/stat' % dev).readline().strip().split(' ') clean = [] for num in file_data: if num != '': clean.append(int(num)) reads[dev].append(clean[0]) writes[dev].append(clean[4]) print reads[dev] print writes[dev] time.sleep(1) print "Device Read Write" print "--------------------------------------" for dev in check_devices: clean_reads = [] reads[dev].reverse() for test, result in enumerate(reads[dev]): if test > 0: clean_reads.append(float(reads[dev][test - 1] - result)) rops = int(math.ceil(sum(clean_reads) / len(clean_reads))) clean_writes = [] writes[dev].reverse() for test, result in enumerate(writes[dev]): if test > 0: clean_writes.append(float(writes[dev][test - 1] - result)) wops = int(math.ceil(sum(clean_writes) / len(clean_writes))) print "%s %s %s" % (dev.ljust(13), repr(rops).ljust(11), repr(wops)) |