From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266778AbUHCSA4 (ORCPT ); Tue, 3 Aug 2004 14:00:56 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266789AbUHCSA4 (ORCPT ); Tue, 3 Aug 2004 14:00:56 -0400 Received: from bay7-f23.bay7.hotmail.com ([64.4.11.23]:21523 "EHLO hotmail.com") by vger.kernel.org with ESMTP id S266778AbUHCSAu (ORCPT ); Tue, 3 Aug 2004 14:00:50 -0400 X-Originating-IP: [193.237.129.167] X-Originating-Email: [royhills@hotmail.com] From: "roy hills" To: linux-kernel@vger.kernel.org Subject: Where is my memory going? (2.4.24 Kernel NFS server) Date: Tue, 03 Aug 2004 19:00:49 +0100 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 03 Aug 2004 18:00:50.0048 (UTC) FILETIME=[CF664800:01C47983] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'm running a Debian "Woody" x86 (P III) Linux system with 2.4.24 kernel which is acting as an NFS file server (using Kernel NFS support) and a PostgreSQL database server. I have 768MB RAM in this system, which is all detected by the Kernel. I can't work out where all the memory is being used, because if I subtract the RSS sizes of all processes (from "ps aux") from 768MB, I get an answer which is _much_ bigger than I get from "free", even after I adjust for buffers and cache. I suspect that the Kernel NFS server is using the memory, but I can't find how to prove this, or how to get the NFS server to report its usage. The nfsd processes show zero RSS in a ps listing. I'm not 100% sure that this is a Kernel issue, so I apologise for this posting if this turns out to be a user-space issue. Here is an example: neon# uname -a Linux neon 2.4.24 #2 Sun Feb 22 12:49:11 GMT 2004 i686 unknown neon# uptime 11:37:21 up 28 days, 17:57, 1 user, load average: 0.06, 0.07, 0.02 neon# free total used free shared buffers cached Mem: 775728 728540 47188 0 2116 6972 -/+ buffers/cache: 719452 56276 Swap: 590952 3232 587720 So we've got about 700MB used and about 55MB free after allowing for buffers and cache. neon:/home/rsh# ps aux | awk 'BEGIN {i=0}; {i+=$6}; END {print i}' 11232 Summing up the RSS fields from a "ps aux" listing shows that the processes are using about 11MB. So where is the other 689MB going? neon:/home/rsh# COLUMNS=75 ps aux USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND root 1 0.0 0.0 1272 72 ? S Jul02 0:03 init [2] root 2 0.0 0.0 0 0 ? SW Jul02 0:00 [keventd] root 3 0.0 0.0 0 0 ? SWN Jul02 0:00 [ksoftirqd_C root 4 0.0 0.0 0 0 ? SW Jul02 24:38 [kswapd] root 5 0.0 0.0 0 0 ? SW Jul02 0:10 [bdflush] root 6 0.0 0.0 0 0 ? SW Jul02 0:04 [kupdated] root 7 0.0 0.0 0 0 ? SW< Jul02 0:00 [mdrecoveryd root 8 0.0 0.0 0 0 ? SW< Jul02 0:00 [raid1d] root 9 0.0 0.0 0 0 ? SW< Jul02 0:00 [raid1d] root 10 0.0 0.0 0 0 ? SW< Jul02 0:00 [raid1d] root 11 0.0 0.0 0 0 ? SW< Jul02 0:00 [raid1d] root 12 0.0 0.0 0 0 ? SW< Jul02 0:00 [raid1d] root 13 0.0 0.0 0 0 ? SW Jul02 0:01 [kjournald] root 102 0.0 0.0 0 0 ? SW Jul02 0:01 [kjournald] root 103 0.0 0.0 0 0 ? SW Jul02 1:43 [kjournald] root 104 0.0 0.0 0 0 ? SW Jul02 4:54 [kjournald] root 105 0.0 0.0 0 0 ? SW Jul02 5:57 [kjournald] root 106 0.0 0.0 0 0 ? SW Jul02 0:01 [kjournald] daemon 128 0.0 0.0 1384 84 ? S Jul02 0:00 /sbin/portma root 224 0.0 0.0 2024 216 ? S Jul02 0:57 /sbin/syslog root 227 0.0 0.0 2032 4 ? S Jul02 0:00 /sbin/klogd root 231 0.0 0.0 1444 4 ? S Jul02 0:00 /sbin/rpc.st root 237 0.0 0.0 1988 4 ? S Jul02 0:00 /usr/sbin/in root 246 0.0 0.0 0 0 ? SW Jul02 0:50 [nfsd] root 247 0.0 0.0 0 0 ? SW Jul02 0:43 [nfsd] root 248 0.0 0.0 0 0 ? SW Jul02 0:50 [nfsd] root 249 0.0 0.0 0 0 ? SW Jul02 0:00 [lockd] root 250 0.0 0.0 0 0 ? SW Jul02 0:00 [rpciod] root 251 0.0 0.0 0 0 ? SW Jul02 0:57 [nfsd] root 252 0.0 0.0 0 0 ? SW Jul02 0:49 [nfsd] root 253 0.0 0.0 0 0 ? SW Jul02 0:52 [nfsd] root 254 0.0 0.0 0 0 ? SW Jul02 0:46 [nfsd] root 255 0.0 0.0 0 0 ? SW Jul02 0:48 [nfsd] root 258 0.0 0.0 1540 4 ? S Jul02 0:00 /usr/sbin/rp mail 261 0.0 0.0 1720 172 ? S Jul02 0:03 /usr/sbin/nu root 316 0.0 0.0 3776 612 ? S Jul02 0:08 /usr/sbin/sn root 322 0.0 0.0 2780 268 ? S Jul02 0:17 /usr/sbin/ss root 331 0.0 0.2 1976 1968 ? SL Jul02 0:00 /usr/sbin/nt daemon 334 0.0 0.0 1384 52 ? S Jul02 0:00 /usr/sbin/at root 337 0.0 0.0 1652 168 ? S Jul02 0:00 /usr/sbin/cr root 344 0.0 0.0 1252 4 tty1 S Jul02 0:00 /sbin/getty root 345 0.0 0.0 1252 4 tty2 S Jul02 0:00 /sbin/getty root 16610 0.0 0.1 6328 1516 ? S 11:31 0:00 /usr/sbin/ss rsh 16612 0.0 0.2 6428 1580 ? S 11:31 0:00 /usr/sbin/ss rsh 16613 0.0 0.1 2220 1264 pts/0 S 11:31 0:00 -bash root 16621 0.0 0.1 2224 1252 pts/0 S 11:32 0:00 bash root 16650 0.0 0.1 3512 1516 pts/0 R 11:38 0:00 ps aux Some other random details: Using software RAID-1 with two IDE disks. Hardware is Dell 500SC server. I'm sure I must be missing something obvious here. Any ideas or pointers? Roy _________________________________________________________________ It's fast, it's easy and it's free. Get MSN Messenger today! http://www.msn.co.uk/messenger