xfs
[Top] [All Lists]

Re: XFS performance issues: O_DIRECT and Linux 2.6.6+

To: james.foris@xxxxxxxxxx
Subject: Re: XFS performance issues: O_DIRECT and Linux 2.6.6+
From: Steve Lord <lord@xxxxxxx>
Date: Tue, 14 Sep 2004 10:15:28 -0500
Cc: Nathan Scott <nathans@xxxxxxx>, linux-xfs@xxxxxxxxxxx
In-reply-to: <414706FA.1040202@xxxxxxxxxx>
References: <411A8410.2030000@xxxxxxxxxx> <20040910041106.GA14336@frodo> <4144B19A.2020407@xxxxxxxxxx> <4145D141.1040907@xxxxxxxxxx> <20040914095914.A4118499@xxxxxxxxxxxxxxxxxxxxxxxx> <414706FA.1040202@xxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mozilla Thunderbird 0.7.1 (X11/20040626)
James Foris wrote:
Nathan Scott wrote:

Hi James,

On Mon, Sep 13, 2004 at 11:56:33AM -0500, James Foris wrote:

More correctly, it happened between 2.6.5 and 2.6.5-bk1

So..... something in the 2.6.5-bk1 patchset caused the change.
Any suggestions where to begin looking (other than fs/direct_io.x) ?



http://kernel.org/pub/linux/kernel/v2.6/snapshots/old/patch-2.6.5-bk1.log


Yup... a bunch more.

"Major surgery against the pagecache, radix-tree and writeback code"

The interesting question is; why do XFS and REISER suffer under O_DIRECT
when other file systems improve?  And why does s/w RAID0 with an external
journal suffer much worse than a simple file system with an internal journal?

Do these questions suggest anyplace else to look ?

Cache invalidation before and after O_DIRECT seems like a good starting
point. The raid0 external journal thing is probably a separate issue.

Steve


<Prev in Thread] Current Thread [Next in Thread>