XFS buffer IO performance is very poor

[Carlos Maiolino]

Hello,

On Fri, Feb 13, 2015 at 10:20:44AM +0800, yy wrote:
>    Dave,
> 
>    Thank you very much for your explanation.
> 
>    I hit this issue when run MySQL on XFS. Direct IO is very import for
>    MySQL on XFS,but I can’t found any document explanation this
>    problem.Maybe this will cause great confusion for other MySQL users
>    also, so maybe  this problem should be explained in XFS document.

I don't think this is something that should be explained in XFS documentation,
but at filesystems documentation in general. Once xfs follows the POSIX
requirements, it's not a "out of standards" behavior, but otherwise, so, I agree
that this should be something documented, but not exactly in XFS itself.

>    Best regards,
>    yy
>     原始邮件
>    发件人: Dave Chinner<david at fromorbit.com>
>    收件人: yy<yy at xspring.net>
>    抄送: xfs<xfs at oss.sgi.com>; Eric Sandeen<sandeen at sandeen.net>;
>    bfoster<bfoster at redhat.com>
>    发送时间: 2015年2月13日(周五) 05:04
>    主题: Re: XFS buffer IO performance is very poor
> On Thu, Feb 12, 2015 at 02:59:52PM +0800, yy wrote:
> > In functionxfs_file_aio_read, will requestXFS_IOLOCK_SHARED lock
> > for both direct IO and buffered IO:
> 
> > so write will prevent read in XFS.
> >
> > However, in function generic_file_aio_read for ext3, will not
> > lockinode-i_mutex, so write will not prevent read in ext3.
> >
> > I think this maybe the reason of poor performance for XFS. I do
> > not know if this is a bug, or design flaws of XFS.
> 
> This is a bug and design flaw in ext3, and most other Linux
> filesystems. Posix states that write() must execute atomically and
> so no concurrent operation that reads or modifies data should should
> see a partial write. The linux page cache doesn't enforce this - a
> read to the same range as a write can return partially written data
> on page granularity, as read/write only serialise on page locks in
> the page cache.
> 
> XFS is the only Linux filesystem that actually follows POSIX
> requirements here - the shared/exclusive locking guarantees that a
> buffer write completes wholly before a read is allowed to access the
> data. There is a down side - you can't run concurrent buffered reads
> and writes to the same file - if you need to do that then that's
> what direct IO is for, and coherency between overlapping reads and
> writes is then the application's problem, not the filesystem...
> 
> Maybe at some point in the future we might address this with ranged
> IO locks, but there really aren't many multithreaded programs that
> hit this issue...
> 
> Cheers,
> 
> Dave.
> --
> Dave Chinner
> [1]david at fromorbit.com
> 
> References
> 
>    1. mailto:david at fromorbit.com

> _______________________________________________
> xfs mailing list
> xfs at oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs


-- 
Carlos