All of lore.kernel.org
 help / color / mirror / Atom feed
From: Arun Sharma <asharma@fb.com>
To: linux-mm@kvack.org
Subject: Accounting for "missing RAM"
Date: Thu, 6 Oct 2011 12:23:31 -0700	[thread overview]
Message-ID: <4E8E0033.70602@fb.com> (raw)


I wrote a script to parse /proc/zoneinfo to figure out how memory on my 
server was getting used.

On many machines this is able to account for RAM +/- 10MB (which I 
consider within the margin of error).

But on many machines, there is 2-3GB of RAM that's unaccounted for and 
mysteriously missing from zoneinfo as well as /proc/meminfo.

When I parse /proc/kpageflags, pages with flags==0 correspond to 
nr_free_pages and pages with flags=KPF_BUDDY correspond to the number of 
"missing ram pages".

As far as I understand, KPF_BUDDY is set only on the first page of a 
higher order page. So if for some reason, we were counting an order=3 
page as 7 pages instead of 8, it'd explain what I am seeing.

  -Arun

#!/usr/bin/env python

import sys, re, string, os

class NestedDict(dict):
     """Implementation of perl's autovivification feature."""
     def __getitem__(self, item):
         try:
             return dict.__getitem__(self, item)
         except KeyError:
             value = self[item] = type(self)()
             return value

sep = re.compile('^Node (.*)')
pages = re.compile('.*nr.*pages.*')
slab = re.compile('.*nr.*slab.*')
present = re.compile('.*present.*')
pcpu_pageset_count = re.compile('.*count:.*')
zone = None
page_size = os.sysconf('SC_PAGESIZE')

zones = NestedDict()
for line in open('/proc/zoneinfo').readlines():
     m = sep.match(line)
     if m:
         zone = m.group(1)
         zones[zone]['pcpu_pageset'] = 0
         # Old kernels don't have this in /proc/zoneinfo
         zones[zone]['nr_free_pages'] = 0
     m1 = pages.match(line)
     m2 = slab.match(line)
     m3 = present.match(line)
     if m1 or m2 or m3:
         name, val = string.split(line)
         val = int(val)
         zones[zone][name] = val
         continue
     m4 = pcpu_pageset_count.match(line)
     if m4:
         name, val = string.split(line)
         val = int(val)
         zones[zone]['pcpu_pageset'] += val

global_diff = 0L
for z in zones.keys():
     allocated = zones[z]['present'] - zones[z]['nr_free_pages']
     if allocated < 0: continue
     print '##############'
     print z
     accounted = 0
     for k in ('nr_file_pages', 'nr_anon_pages',
               'nr_slab_reclaimable', 'nr_slab_unreclaimable',
               'nr_page_table_pages', 'pcpu_pageset'):
         val = zones[z][k]
         print k, val
         accounted += val
     print "allocated", "accounted", "diff"
     print allocated, accounted, allocated - accounted
     global_diff += (allocated - accounted)

vmalloc = 0L
vmalloc_re = re.compile('.*VmallocUsed:\s+(\d+).*')
for line in open('/proc/meminfo').readlines():
     m = vmalloc_re.match(line)
     if m:
         vmalloc_kb = int(m.group(1))
vmalloc = vmalloc_kb * 1024/page_size

print '##############'
print 'vmalloc',  vmalloc
print "missing ram: ", (global_diff - vmalloc) * page_size

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

                 reply	other threads:[~2011-10-06 19:23 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4E8E0033.70602@fb.com \
    --to=asharma@fb.com \
    --cc=linux-mm@kvack.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.