All of lore.kernel.org
 help / color / mirror / Atom feed
From: Vladimir Zidar <vladimir@mindnever.org>
To: linux-kernel@vger.kernel.org
Subject: /proc/<pid>/status: VmSize
Date: Fri, 28 Nov 2003 20:18:08 +0100	[thread overview]
Message-ID: <1070047087.4058.469.camel@mravojed> (raw)

 Hola,

 We are running kernel 2.4.22 on i686 SMP box.
Inside are two Intel(R) XEON(TM) CPU 2.00GHz, with hyperthreading
enabled, so /proc/cpuinfo shows four CPUs. 
 This box has 4GB of RAM installed, and 2GB of swap space.

 The problems we are experiencing are related to heavy usage of VM.

 We have multithreaded application that uses lots of mmap()-ed memory
(one big file in size around 700 MB, and lots of anonymous mappings to
/dev/zero in average size between 64k and 1MB). This program sometimes
grow up to 1.6 GB in size (SIZE that is shown by top utility).. But,
sometimes /proc/<pid>/status shows VmSize with more than 2GB, where at
the same time top and other /proc/<pid> entries show 1.6GB. This somehow
affects pthread_create() call which fails then.

 The question is, how can happen that different numbers are shown in
proc filesystem, for same pid ? (which is part of multithreaded
process), and why pthread_create() fails ? Is there maybe 2GB limit on
memory size that single process can manage on i386 ? Also, when such
program crashes, it creates 2GB core file, which is not completly usable
from gdb. (gdb complains that some addresses are not accessible)..
 I suspect that this has something to do with amount of RAM (4GB), but
we are still trying to get this server tested with only 2GB running in
standard (not paged) mode.. but this can take some time, since it is one
of our production machines.


 Anybody, idea ?

 Thanks.


-- 


             reply	other threads:[~2003-11-28 19:13 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-11-28 19:18 Vladimir Zidar [this message]
2003-11-28 19:41 ` /proc/<pid>/status: VmSize Vladimir Zidar
2003-11-28 21:38 ` William Lee Irwin III
2003-12-06 17:57   ` Marcelo Tosatti

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=1070047087.4058.469.camel@mravojed \
    --to=vladimir@mindnever.org \
    --cc=linux-kernel@vger.kernel.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.