Re: [PATCH 0/3 v5][RFC] ext3/4: enhance fsync performance when using CFQ

From: Tao Ma <tao.ma@oracle.com>
To: Jeff Moyer <jmoyer@redhat.com>
Cc: axboe@kernel.dk, vgoyal@redhat.com, linux-kernel@vger.kernel.org,
	linux-ext4@vger.kernel.org
Subject: Re: [PATCH 0/3 v5][RFC] ext3/4: enhance fsync performance when using CFQ
Date: Wed, 23 Jun 2010 17:30:42 +0800	[thread overview]
Message-ID: <4C21D442.8080703@oracle.com> (raw)
In-Reply-To: <1277242502-9047-1-git-send-email-jmoyer@redhat.com>

Hi Jeff,

On 06/23/2010 05:34 AM, Jeff Moyer wrote:
> Hi,
>
> Running iozone with the fsync flag, or fs_mark, the performance of CFQ is
> far worse than that of deadline for enterprise class storage when dealing
> with file sizes of 8MB or less.  I used the following command line as a
> representative test case:
>
>    fs_mark -S 1 -D 10000 -N 100000 -d /mnt/test/fs_mark -s 65536 -t 1 -w 4096 -F
>
> When run using the deadline I/O scheduler, an average of the first 5 numbers
> will give you 448.4 files / second.  CFQ will yield only 106.7.  With
> this patch series applied (and the two patches I sent yesterday), CFQ now
> achieves 462.5 files / second.
which 2 patches? Could you paste the link or the subject? Just want to 
make my test env like yours. ;)
As Joel mentioned in another mail, ocfs2 also use jbd/jbd2, so I'd like 
to give it a try and give you some feedback about the test.

Regards,
Tao
>
> This patch set is still an RFC.  I'd like to make it perform better when
> there is a competing sequential reader present.  For now, I've addressed
> the concerns voiced about the previous posting.
>
> Review and testing would be greatly appreciated.
>
> Thanks!
> Jeff
>
> ---
>
> New from the last round:
>
> - removed the think time calculation I added for the sync-noidle service tree
> - replaced above with a suggestion from Vivek to only guard against currently
>    active sequential readers when determining if we can preempt the sync-noidle
>    service tree.
> - bug fixes
>
> Over all, I think it's simpler now thanks to the suggestions from Jens and
> Vivek.
>
> [PATCH 1/3] block: Implement a blk_yield function to voluntarily give up the I/O scheduler.
> [PATCH 2/3] jbd: yield the device queue when waiting for commits
> [PATCH 3/3] jbd2: yield the device queue when waiting for journal commits
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>