xfs
[Top] [All Lists]

Re: Speed up xfsdump ?

To: linux-xfs@xxxxxxxxxxx
Subject: Re: Speed up xfsdump ?
From: Nicolas Kowalski <Nicolas.Kowalski@xxxxxxx>
Date: Tue, 13 Sep 2005 18:48:02 +0200
In-reply-to: <Pine.LNX.4.63.0509130933130.25299@twinlark.arctic.org>
References: <4325BA69.4070709@sgi.com> <20050912174223.51325.qmail@web34104.mail.mud.yahoo.com> <20050913123852.Z4874818@boing.melbourne.sgi.com> <Pine.LNX.4.63.0509131303010.6178@corbeau.imag.fr> <4326D34C.9010404@sgi.com> <Pine.LNX.4.63.0509131541410.6974@corbeau.imag.fr> <Pine.LNX.4.63.0509130933130.25299@twinlark.arctic.org>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Tue, 13 Sep 2005, dean gaudet wrote:

> On Tue, 13 Sep 2005, Nicolas Kowalski wrote:
> 
> > PS: as an additional test, on the same array/server, I created an ext3 
> >     filesystem, pushed the same files than for the XFS/xfsdump 
> >     tests, and ran a dump session:
> 
> if the xfs filesystem was created "live" and had files added/deleted 
> along the way it's likely to be more fragmented than a freshly created 
> ext3 filesystem which had a bunch of files copied to it right after 
> being created... one more test you might want to do is a fresh xfs fs 
> which you populate the same way you populated the ext3 fs.

That's what I do: all my tests with xfsdump are done on a newly 
created/populated filesystem, so that fragmentation does not "corrupt" 
my results.

-- 
Nicolas


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