计算内存和CPU利用率

Question 1

ps是最简单的信息接口/proc。

以下是列出每个用户内存的一种方法：

$ ps -e -o uid,vsz | awk '
{ usage[$1] += $2 }
END { for (uid in usage) { print uid, ":", usage[uid] } }'

如果你真的想使用 proc，我建议使用 Python 或 Perl 之类的东西来迭代一次/proc/*/status，并将用户/使用键/值对存储在哈希中。

的相关字段/proc/PID/status似乎是：

Uid:    500     500     500     500
VmSize:     1234 kB

我认为这四个 Uid 数字是真实 uid、有效 uid、保存 uid 和 fs uid。

假设你想要真正的 uid，类似这样的东西应该可以工作：

# print uid and the total memory (including virtual memory) in use by that user
# TODO add error handling, e.g. not Linux, values not in kB, values not ints, etc.

import os
import sys
import glob

# uid=>vsz in KB
usermem = {}

# obtain information from Linux /proc file system
# http://www.kernel.org/doc/man-pages/online/pages/man5/proc.5.html
os.chdir('/proc')
for file in glob.glob('[0-9]*'):
    with open(os.path.join(file, 'status')) as status:
        uid = None
        mem = None
        for line in status:
            if line.startswith('Uid:'):
                label, ruid, euid, suid, fsuid = line.split()
                uid = int(ruid)
            elif line.startswith('VmSize:'):
                label, value, units = line.split()
                mem = int(value)
        if uid and mem:
            if uid not in usermem:
                usermem[uid] = 0
            usermem[uid] += mem

for uid in usermem:
    print '%d:%d' % (uid,usermem[uid])

CPU就更难了。

ps(1) 手册页显示：

   CPU usage is currently expressed as the percentage of time spent
   running during the entire lifetime of a process. This is not ideal,
   and it does not conform to the standards that ps otherwise conforms to.
   CPU usage is unlikely to add up to exactly 100%.

所以我不确定。也许看看top它是如何处理的。或者你可以ps -e -o uid,pid,elapsed在给定的时间间隔运行两次，然后减去两次，或者其他什么。

或者安装更适合此目的的东西，例如流程会计。

Answer