Date: Fri, 29 Jun 2018 12:00:12 -0600 From: "@lbutlr" <kremels@kreme.com> To: freebsd-questions <freebsd-questions@freebsd.org> Subject: Re: swap_pager_getswapspace Message-ID: <8D5A0F22-4017-4A37-843E-07931CE94D32@kreme.com> In-Reply-To: <20180629202603.0492ab82.freebsd.ed.lists@sumeritec.com> References: <FD06A58F-27A5-4B25-8D67-B85620268A8C@kreme.com> <20180629202603.0492ab82.freebsd.ed.lists@sumeritec.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 29 Jun 2018, at 06:26, Erich Dollansky <freebsd.ed.lists@sumeritec.com> wrote:
> there was recently a long discussion regarding this. My solution is to
> create a file-backed memory disk and use it for swap when the media is
> not able or has to be reformatted. I still use the old-fashioned 2 *
> RAM size to determine my swap space. It seems overkill but I never had
> problems with this saving a lot of my time.
Top, after killing off some processes:
Mem: 1204M Active, 424M Inact, 1072K Laundry, 219M Wired, 85M Buf, 1542M Free
It looks like what is happening is that many rsync daemon processes are bing created and do not die. I just checked and there were 49 running and nearly no free memory (17M). after a killall rysnc the free memory jumped to 1542M and left 9 processes still running for several minutes. By the time those died, memory was up to 2300M free. So it appears the issue is less swap and more rsync spawning many processes for some reason.
# rsyncd.log
# Set this if you want to stop rsync daemon with rc.d scripts
pid file = /var/run/rsyncd.pid
uid = root
gid = wheel
[root]
path = /
auth users = root
secrets file = /usr/local/etc/rsyncd.secrets
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8D5A0F22-4017-4A37-843E-07931CE94D32>