----- Original Message -----
From: "Jan Derfinak" <ja@xxxxxxxxxxxx>
To: "Dmitry Melekhov" <dm@xxxxxxxxxx>
Cc: "XFS Mailing List" <linux-xfs@xxxxxxxxxxx>
Sent: Thursday, April 24, 2003 10:54 AM
Subject: Re: filesystem size difference ? ;-)
> On Wed, 23 Apr 2003, Dmitry Melekhov wrote:
> > Only question - is this difference between Suse xfs ;-) and SGI xfs or
> > between 2.4.19 and 2.4.20.
> SuSE has XFS 1.1 in SLES8. And kernel from SuSE has lot of other patches
> which isn't in vanilla kernel. If you think that differnce is in kernel
> to boot your 2.4.20 kernel in SLES8 and repeat copying.
Look like you are right, problem is in Suse kernel.
Here is result of
xfsdump -l 0 - /home/samba/common | rsh vader "/sbin/xfsrestore -
on old server:
[root@p100 common]# df |grep common
/dev/vg02/lvcommon 9432384 8656128 776256 92% /home/samba/common
on new server:
vader:/home/samba/common # df |grep common
/dev/vga3/lvcommon 9432384 8655244 777140 92% /home/samba/common
Difference in used blocks can be easily explained- I don't know why
xfs_restore can't set acls ( I wrote mail
about this problem to list today), but I setted default acls on new server
(vader) similar to which I have
on old server (p100). Unfortunately ;-) some files have acls different from
default. I'll fix this tomorrow...
Certanly, I'll write to Suse mail list about this problem...
btw, if have another question- is it possible to move info from one server
using dd ? I want to
have logic volumes with the same size... And resolving groups and users on
old server is too slow :-( ,
so star takes several hours to copy files, xfsdump works far fatser, but it
can't set acls...