Date: Wed, 30 Mar 2011 21:46:13 -0500 From: Adam Vande More <amvandemore@gmail.com> To: Luke Dean <LukeD@pobox.com> Cc: freebsd-questions@freebsd.org Subject: Re: kern.geom.journal.stats.low_mem refers to what? Message-ID: <AANLkTim5YYHObTzq-VTZb7CfwiBhV-H-AjdeB2FZDtB=@mail.gmail.com> In-Reply-To: <alpine.BSF.2.00.1103301830050.36948@tower.lukas.is-a-geek.org> References: <alpine.BSF.2.00.1103271901080.16486@tower.lukas.is-a-geek.org> <alpine.BSF.2.00.1103301830050.36948@tower.lukas.is-a-geek.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Mar 30, 2011 at 9:07 PM, Luke Dean <LukeD@pobox.com> wrote: > > The Handbook says 1GB is good enough most of the time, but it also says >> that 3x the amount of physical memory is a good size as well. I compromised >> between the two and made an 8GB journal for this system that has ~4GB of >> memory. >> > > I've never understood why that article states that calculation. The amount of RAM a system has little to do with what gjournal needs. That formula will in most cases allocate far more journal space than is really necessary. Disk speed and latency are far more important factors. Back when I was using gjournal, I used a 2GB journal on 7200rpm and was never able to panic the system unlike with a 1 GB journal but that was with all other settings at default. Allocating more to gjournal cache would increase the need for space so perhaps it's good you've made such a large one. As you've discovered, gjournal's performance isn't so great when large files are involved. It still helps on multi-threaded access but those big sequential operations destroy most the other optimizations in place. You might want to consider placing your journal on an SSD. -- Adam Vande More
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTim5YYHObTzq-VTZb7CfwiBhV-H-AjdeB2FZDtB=>