All of lore.kernel.org
 help / color / mirror / Atom feed
From: Brice Figureau <brice+lklm@daysofwonder.com>
To: Chris Mason <chris.mason@oracle.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: Sluggish system while copying large files.
Date: Mon, 23 Oct 2006 11:40:05 +0200	[thread overview]
Message-ID: <1161596405.473.29.camel@localhost.localdomain> (raw)
In-Reply-To: <20061018150901.GC16570@think.oraclecorp.com>

Hi,

On Wed, 2006-10-18 at 11:09 -0400, Chris Mason wrote:
> On Fri, Oct 13, 2006 at 03:56:14PM +0200, Brice Figureau wrote:
> > I have a brand new Dell 2850 biXeon x86_64 with a Perc4e/Di (megaraid)
> > RAID card with two hardware RAID1 volumes (sda and sdb, ext3 on top of
> > LVM2, io scheduler deadline).
> > 
> > This machine runs 2.6.18 and is used as a mysql server.
> > 
> > Whenever I cp large files (for instance during backup) from one volume
> > to the other, the system becomes really sluggish. 
> > [snipped my long explanation] 
> > Mysql data is on sdb. The copy takes place from sdb to sda.
> 
> The first thing I would suggest would be to use data=writeback for the
> mysql partitions.  The easiest way to figure out what is causing the
> latencies is to get the output from sysrq-t during one of the stalls.
> 
> (see Documentation/sysrq.txt for details on enabling sysrq).

Unfortunately, sysrq-t has an unfortunate tendency to freeze the box
because of a problem in the stack unwinder in this kernel version (as I
read somewhere on lklm).
Last time I tried, it locked the box and I had some problem to revive it
(especially because there was a running LVM snapshot that wasn't
deactivated and the next reboot stalled, I still have to check why).

I'll try the data=writeback mount option. Is it possible to remount the
partition and change the journal option on a running system?

> Since mysql is probably triggering tons of fsyncs or O_SYNC writes,
> you may want to increase the size of the ext3 log.

Mysql is using O_DIRECT for its datafile. I don't know how it relates to
the sync things, but I guess that to be truly ACID, it has to fsync the
files on each transactions.
How can I increase the ext3 log ?
Any idea of the size I should use (and what is the default) ?

> If mysql is constantly appending to the files holding your tables, the
> synchronous writes are more expensive and log intensive.  Check your
> setup to see if you can manually extend any of those files to avoid
> constantly growing table files.

The use of a battery-backed RAID cache should mitigate the sync writes,
and since our mysql load is quite low, the machine shouldn't definitely
freeze for seconds while copying files.

Last week I opened bug #7372 on bugzilla.kernel.org where I gave more
information about this problem.

Thank you for your help,
-- 
Brice Figureau
Days of Wonder


  reply	other threads:[~2006-10-23  9:40 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-10-13 13:56 Sluggish system while copying large files Brice Figureau
2006-10-13 14:14 ` Paolo Ornati
2006-10-18 15:09 ` Chris Mason
2006-10-23  9:40   ` Brice Figureau [this message]
2006-10-23 11:29     ` Chris Mason

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1161596405.473.29.camel@localhost.localdomain \
    --to=brice+lklm@daysofwonder.com \
    --cc=chris.mason@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.