Date: Thu, 15 Sep 2016 16:10:26 +0100 From: RW <rwmaillists@googlemail.com> To: freebsd-questions@freebsd.org Subject: Re: Best kind of hard drive for heavy use? Message-ID: <20160915161026.62dffff7@gumby.homeunix.com> In-Reply-To: <20160915155254.768f6f70@archlinux.localdomain> References: <42.56.05022.D3A48D75@dnvrco-oedge02> <20160914120349.76a015cd@gumby.homeunix.com> <20160914175449.185d12b0@archlinux.localdomain> <20160914221954.00fb1d56@gumby.homeunix.com> <20160915013848.5564c238@archlinux.localdomain> <CAOyJeZTzo4Kh9OaKQk6_-6qB8imHbGGMgT53DNK0%2BNgS-HR37g@mail.gmail.com> <20160915140856.24af27ca@gumby.homeunix.com> <20160915155254.768f6f70@archlinux.localdomain>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 15 Sep 2016 15:52:54 +0200 Ralf Mardorf via freebsd-questions wrote: > On Thu, 15 Sep 2016 14:08:56 +0100, RW via freebsd-questions wrote: > >There's no such thing, you can specify the maximum size of a tmpfs, > >but not how much is kept in RAM. It doesn't matter though, because > >the VM system will handle it better than a simple limit. > > That might be true, but not necessarily is true. It is if you're using UFS and set a sensible tmpfs limit. ZFS does complicate this because it's not well integrated into VM, and I think it still has some problems with swap. > If the handling would > be that good, then I wonder why on Linux the default without an fstab > entry usually is half of the RAM and not the complete RAM. That default doesn't take account of swap, and in the absence of swap you cannot use 100% of the RAM for tmpfs. The important thing is that everything that can't (or shouldn't) be discarded has to fit into swap+ram. Most desktops/workstations have much more memory than they need, in which case you can safely allocate swap plus most of the ram to tmpfs, if you want to.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160915161026.62dffff7>