Date: Mon, 16 Jul 2001 10:55:13 -0700 From: David Greenman <dg@root.com> To: Rob <rob@breakbeat.net> Cc: freebsd-questions@freebsd.org Subject: Re: Inactive/leak memory Message-ID: <20010716105513.G49840@nexus.root.com> In-Reply-To: <Pine.BSF.4.21.0107161837370.1233-100000@phoenix.shells.co.uk>; from rob@breakbeat.net on Mon, Jul 16, 2001 at 06:49:46PM %2B0100 References: <Pine.BSF.4.21.0107161837370.1233-100000@phoenix.shells.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
>Hi, > >I'm running 4.3 on a dual P3-550, with 1Gb RAM, and have been noticing odd >reports in the 'top' values. After a reboot, the server slowly begins to lose >memory from the free column into the inactive column, resulting in a bizarre set >of values. >This server is a live webserver, running Apache, with PHP and PostgreSQL. The >laods are fairly low, and since I've had this box constantly CVSupped, with the >latest updates to the webserver etc, I'm pretty stumped as to why this situation >is happening. > >Here's the top part of the top output, after only 15 days of uptime: > >last pid: 99471; load averages: 0.56, 0.24, 0.13 > up 15+21:26:00 17:13:02 >50 processes: 1 running, 49 sleeping >CPU states: 1.9% user, 0.0% nice, 1.4% system, 0.4% interrupt, 96.3% idle >Mem: 22M Active, 668M Inact, 87M Wired, 152K Cache, 112M Buf, 227M Free >Swap: 512M Total, 512M Free > > PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU COMMAND > 189 pgsql 2 0 4408K 1552K select 1 11:36 0.00% 0.00% postgres > 207 root 2 0 4856K 4012K select 1 4:43 0.00% 0.00% httpd >72058 www 2 0 5252K 4624K sbwait 1 0:55 0.00% 0.00% httpd >79544 www 18 0 5300K 4584K lockf 0 0:35 0.00% 0.00% httpd >80731 www 2 0 5184K 4488K sbwait 1 0:35 0.00% 0.00% httpd >80748 www 2 0 5168K 4536K sbwait 1 0:35 0.00% 0.00% httpd >85255 www 18 0 5176K 4544K lockf 0 0:26 0.00% 0.00% httpd > 154 root 2 0 1256K 884K select 1 0:26 0.00% 0.00% sshd1 > 147 root 10 0 968K 736K nanslp 0 0:07 0.00% 0.00% cron > 150 root 2 0 2484K 2036K select 0 0:06 0.00% 0.00% sendmail > 124 root 2 0 924K 612K select 1 0:05 0.00% 0.00% syslogd >95582 www 2 0 5244K 4604K sbwait 1 0:04 0.00% 0.00% httpd >95470 www 2 0 5068K 4368K sbwait 1 0:04 0.00% 0.00% httpd >98534 www 2 0 5076K 4432K sbwait 0 0:01 0.00% 0.00% httpd > 194 root 2 0 1236K 888K select 0 0:01 0.00% 0.00% sshd1 > >This seems to happen slowly, until it reaches about 20M Free, upon which I tend >to reboot. > >Any ideas? I thought that perhaps as inactive memory is still available to the >system if needed that it might be okay to leave it, but a explanation may put my >heart at rest.. The active, inactive, cache, and free above are actually page queues, and not an indication of actual memory usage. The system moves the pages between the queues as part of the page reclaimation procedure. The 'free' pages are the easiest to reclaim because they contain no valid data. The 'free' number will generally be *low* because FreeBSD tries to keep pages filled with useful cached file data (and those pages will show up on any of the other queues depending on how/when they were last accessed). The bottom line is that you can't determine anything about truely allocated memory by looking at those page queue numbers. -DG David Greenman Co-founder, The FreeBSD Project - http://www.freebsd.org President, TeraSolutions, Inc. - http://www.terasolutions.com Pave the road of life with opportunities. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010716105513.G49840>