xfs
[Top] [All Lists]

Re: mount: Function not implemented?

To: Anthony Biacco <ABiacco@xxxxxxxxx>
Subject: Re: mount: Function not implemented?
From: Chris Wedgwood <cw@xxxxxxxx>
Date: Mon, 19 Jul 2004 15:57:25 -0700
Cc: Steve Lord <lord@xxxxxxx>, linux-xfs@xxxxxxxxxxx
In-reply-to: <74918D8CA17F7C418753F01078F10B6BD08560@bill.corporate.quris.com>
References: <74918D8CA17F7C418753F01078F10B6BD08560@bill.corporate.quris.com>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Mon, Jul 19, 2004 at 04:50:05PM -0600, Anthony Biacco wrote:

> But it's a 64-bit system.

with a crappy MMU :( that's by far my biggest bitch with x86-64

> How do people get oracle performing on enterprise class hardware,
> with linux, with such a low page size?

i really don't think you'll see much performance difference between 4k
and 16k pages (on a cpu that does allow this)

> Do you just have to say, the hell with it, and create a raw device?

you can try that, but i don't think you'll see a significant
performance difference using 4k blocks

linux doesn't do IO in the size of the fs' blocksize, raw or otherwise


  --cw


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