xfs
[Top] [All Lists]

Re: re[2]: Redhat Advanced Server and XFS?

To: Greg Freemyer <freemyer@xxxxxxxxxxxxxxxxx>
Subject: Re: re[2]: Redhat Advanced Server and XFS?
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Thu, 22 Aug 2002 18:45:25 +0100
Cc: Simon Matter <simon.matter@xxxxxxxxxxxxxxxx>, Bill Anderson <bill@xxxxxxxxxxx>, Linux XFS <linux-xfs@xxxxxxxxxxx>
In-reply-to: <20020822165802.ZFAJ25788.imf08bis.bellsouth.net@TAZ2>; from freemyer@NorcrossGroup.com on Thu, Aug 22, 2002 at 12:55:13PM -0400
References: <20020822165802.ZFAJ25788.imf08bis.bellsouth.net@TAZ2>
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Mutt/1.2.5.1i
On Thu, Aug 22, 2002 at 12:55:13PM -0400, Greg Freemyer wrote:
> Per the interview at 
> http://ocfs.otncast.otnxchange.oracle.com/source/browse/ocfs/
>  Oracle has added a lot of Enterprise quality tweaks to the AS kernel.  (I 
> think many of these are now in 2.4.19, but I'm not sure.)

"Enterprise quality tweaks" is exactly the right term.  Meaning they made
Ingo code up some crude hacks to win benchmarks. (and some useful stuff
like the O(1) scheduler).

It's not that dramatic as Orcale describes it.  And no, I haven't looked at
that page as I refuse the silly registration.  If Oracle wants to cooperate
with Linux developers they better make their information freely available
without having to sign up for spam.

> If you are considering a Production Quality Oracle environment, you probably 
> need to stick to the AS kernel.  

you have to anyway when you want to get their support.  This has nothing to
do with the questionable quality of that kernel, but rather with Oracle's
ceritifcation policy.

If you want to chose your kernel run DB/2 or some opensouce database, it's
pretty simple..

> [Off Topic below]  
> Christoph: Do you know anything about the newly GPLed Oracle Cluster File 
> System (OCFS)?

I can't find their code anyway, but I've been told that it is enteprise grade 
code
(i.e. horrible code quality but fixable).

> I'm hoping it might fill the role that OpenGFS is currently filling. (Alan 
> Cox has not had good things to say about OpenGFS, so I assume it is not a 
> long term player.)

I've worked on OpenGFS, and I thuink the major problem is that there is no
founding for a badly needed rewrite.  SGI might have an interesting offer
long-term, btw :)


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