xfs
[Top] [All Lists]

Re: changing xfs file system tunnings(sunit,swidth) after mkfs.xfs

To: "Raz Ben-Jehuda(caro)" <raziebe@xxxxxxxxx>
Subject: Re: changing xfs file system tunnings(sunit,swidth) after mkfs.xfs
From: David Chinner <dgc@xxxxxxx>
Date: Tue, 27 Mar 2007 23:20:47 +1100
Cc: nscott@xxxxxxxxxx, xfs@xxxxxxxxxxx
In-reply-to: <5d96567b0703270004j72d89618xe27bbf3d5e44eea1@mail.gmail.com>
References: <5d96567b0703261325m8f17b2eg55e04264fee4832a@mail.gmail.com> <1174947030.5051.586.camel@edge> <5d96567b0703270004j72d89618xe27bbf3d5e44eea1@mail.gmail.com>
Sender: xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4.2.1i
On Tue, Mar 27, 2007 at 09:04:33AM +0200, Raz Ben-Jehuda(caro) wrote:
> ahha...
> but did you ever tried to set an sunit value bigger than the one
> set during the mkfs.xfs ?

Sure - but you've got to satisfy a whole heap of alignment
and size considerations when changing sunit e.g. AGs must be
stripe unit aligned and their size must be a whole multiple
of the stripe unit.

Hence any new value generally needs to be a multiple so
that filesystem alignment remains valid.

You might want to get a recent kernel, too, as it will
tell more about the reason why the mount with the new
sunit is failing (in dmesg).

Cheers,

dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group


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