Date: Tue, 28 Oct 2014 13:01:46 +0000 From: RW <rwmaillists@googlemail.com> To: freebsd-questions@freebsd.org Subject: Re: out of swap space Message-ID: <20141028130146.6d2b6179@gumby.homeunix.com> In-Reply-To: <CA%2BtpaK20Pt0A-G7KzJPhqXn%2BRruk8hF4B4nyhc0uKwxuHvfaMQ@mail.gmail.com> References: <0fdf2022075b7a33f0abde4edd7c12a1@paz.bz> <CA%2BtpaK20Pt0A-G7KzJPhqXn%2BRruk8hF4B4nyhc0uKwxuHvfaMQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 27 Oct 2014 16:39:52 -0500 Adam Vande More wrote: > On Mon, Oct 27, 2014 at 2:57 PM, Jim Pazarena <fquest@paz.bz> wrote: > > > There is a lot of historical chatter about the amount of swap space > > required. > > But for my question, I haven't seen discussion: > > What HAPPENS when the system flags "out of swap space". > > Does a process die? or does the system merely become very sluggish? > > > > Both, a process is killed and whenever you're starting to use swap > space you should expect the system to become sluggish. I don't know that anything bad happens simply because you run out of swap. Processes are killed when the system is unable to find enough memory to carry on. For example if you have 4GB of RAM and 1GB of swap, and you leave 2GB on tmpfs, you may fill swap without even coming close to running out of memory. Another example is a very slow memory leak where running out of memory could happen a long time after running out of swap. There seems to be a common trend of allocating swap space that's much smaller than RAM. When you combine that with tmpfs use, I suspect it may have become much more common to run out of swap without consequences.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20141028130146.6d2b6179>