Date: Sun, 11 May 2014 08:50:35 +0200 From: Gary Jennejohn <gljennjohn@gmail.com> To: svn-src-all@freebsd.org Subject: Re: svn commit: r265418 - head/sys/vm Message-ID: <20140511085035.2a38b298@ernst.home> In-Reply-To: <20140510190734.2ce803313362e16d7af7ca86@gmail.com> References: <201405060342.s463g5Fx047447@svn.freebsd.org> <1399379941.22079.263.camel@revolution.hippie.lan> <5369777F.6010203@rice.edu> <20140510190734.2ce803313362e16d7af7ca86@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 10 May 2014 19:07:34 +0200 Marc UBM <ubm.freebsd@googlemail.com> wrote: > On Tue, 06 May 2014 18:59:59 -0500 > Alan Cox <alc@rice.edu> wrote: > > > > Does this address the observation that several folks have made on > > > current@ that pages are being pushed to swap much more agressively than > > > in the past, even when the system doesn't seem short of memory? > > > > > > > I hope so. Please let me know if you see any difference. > > > > I also suspect that another side-effect of r254304 is that we are > > swapping out idle processes sooner, before memory has become scarce. > > Has anyone observed this behavior? > > This fixes the aggressive swapping I have observed on my machine (used > swap at 50% with 1GB free memory). > > Thanks a lot! > I also saw some improvement in swapping when there is lots of disk activity. My box has 8GB of RAM. I frequently mv lots of ca. 4GB files from one disk to another. What I observe in top is that inactive memory is on the order of 6GB and free memory is on the order of 50MB when the box starts to swap. Shortly after the swap free memory jumps to about 3GB. Seems to me that such a large amount of inactive memory should be taken advantage of to replenish free memory without having to swap out a single byte. Still, the total number of bytes swapped out is considerably less than before this change was made, so there's definitely an improvement. -- Gary Jennejohn
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140511085035.2a38b298>