Date: Wed, 22 Jun 2011 13:18:25 +0200 From: Mikael Fridh <frimik@gmail.com> To: Wiktor Niesiobedzki <bsd@vink.pl> Cc: freebsd-fs@freebsd.org Subject: Re: ZFS L2ARC hit ratio Message-ID: <BANLkTik_575fA9f6eb4B619LT9_qDLSaCA@mail.gmail.com> In-Reply-To: <BANLkTikmtOwYYLjsQBMBND9L6YsLkD8jhg@mail.gmail.com> References: <BANLkTinyveD7D=PYv3eqdxZb=KneKxg9Zg@mail.gmail.com> <BANLkTindsp-WqoTySZtym8LtX7DnMgjD-g@mail.gmail.com> <BANLkTikmtOwYYLjsQBMBND9L6YsLkD8jhg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Jun 22, 2011 at 6:51 AM, Wiktor Niesiobedzki <bsd@vink.pl> wrote: > 2011/6/22 Artem Belevich <art@freebsd.org>: >> >> L2ARC is filled with items evicted from ARC. The catch is that L2ARC >> writes are intentionally throttled. When L2ARC is empty writes happen >> at a higher rate, but it's still intentionally low so that >> read-optimized cache device does not wear out too soon. The bottom >> line is that not all the data spilled out of ARC ends up in L2ARC on >> the first try. Re-run your experiment again and you would probably see >> some improvement in L2ARC hit rates. > > I've run the experiment 3 times with no extent. Funny thing is: > - in first run, I see a lot of write activity against cache device > - in second run, I see no write activity against cache device, nor read activity What about read activity from vdevs? > So my guess is, that anyhow, ZFS cache layer knows, that this file is > *there*, though it decides not to serve it from L2ARC... Have you considered, that it is mostly served from ARC? Or even from the underlying vdevs? How much RAM have you got? And thus; How big is your ARC :) -- Mikael
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?BANLkTik_575fA9f6eb4B619LT9_qDLSaCA>