Date: Fri, 30 Apr 2010 07:14:26 -0600 From: "Jason J. W. Williams" <jasonjwwilliams@gmail.com> To: Pierre Massat <depaulou@efrei.fr> Cc: "freebsd-current@freebsd.org" <freebsd-current@freebsd.org> Subject: Re: ZFS Stability & MySQL (Comments Requested) Message-ID: <80219502-EDAD-4745-A89B-813601D1C15E@gmail.com> In-Reply-To: <p2i4c3c61161004300104iff071a88w2723795ac94cbf76@mail.gmail.com> References: <u2q3b949f091004290731rd83bd474g9b7b2cacf5120b8@mail.gmail.com> <p2i4c3c61161004300104iff071a88w2723795ac94cbf76@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Pierre, What was the iowait when the update hangs occurred? Also, what was the box's uptime approx when it happened? Have you surpassed that amount of uptime since? Were you able to DTrace MySQL to see what func was hanging? Thank you so much! -J Sent via iPhone Is your e-mail Premiere? On Apr 30, 2010, at 2:04, Pierre Massat <depaulou@efrei.fr> wrote: > Hi everybody, > > On Thu, Apr 29, 2010 at 4:31 PM, Jason J. W. Williams > <jasonjwwilliams@gmail.com> wrote: >> I've written before that we're considering moving to FreeBSD 8 from >> OpenSolaris and are heavily reliant on ZFS. Has anyone used FreeBSDs >> ZFS implementation for a high reliability environment like a >> database? >> If so, what are your experiences? > > I'm running MySQL 5.5 (Server version: 5.5.1-m2-log FreeBSD port: > mysql-server-5.5.1) on FreeBSD 8.0 with a quite recent kernel (FreeBSD > 8.0-STABLE (GENERIC) #0: Tue Feb 23 12:38:58 UTC 2010). I've done some > tweaks to ZFS, like limiting ARC size, followed advices like > recordsize matching InnoDB page size (i'm using InnoDB on my 4Go > database). We have 11% of reads and 89% of writes and we make around > 300 queries per second. > > The thing is it runs well except that once, I had a very big slowdown > on MySQL : UPDATE queries were taking 30 seconds (which is actually > the timeout in my.cnf) to complete. I can't tell if it was a MySQL > problem (as i'm using something which is not release yet i think) or > if it's due to ZFS. However, after a restart, everything was back to > normal and things are running well. We have a lot of memory, 32 Go > (yes, the box is actually oversize for what we're doing right now), > and we don't stress the server so much (it has 2x XeonQuand X55xx), > and that might be a reason why it runs well (we don't push ZFS and the > server to the limits). > > I hope it's useful for you. If you need anything more, please ask. > > -- > Pierre Massat
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?80219502-EDAD-4745-A89B-813601D1C15E>