question: are the metadata crc elements ready for general use?

question: are the metadata crc elements ready for general use?

[Joe Landman]

I ask as I see this in the 3.2.2 mkfs.xfs command.  The default is to 
leave them off from what I can see on an mkfs.xfs.

root@unison:~/xfsprogs# mkfs.xfs /dev/sda
meta-data=/dev/sda               isize=256    agcount=55, 
agsize=268435455 blks
          =                       sectsz=512   attr=2, projid32bit=1
          =                       crc=0        finobt=0
data     =                       bsize=4096   blocks=14627632640, imaxpct=1
          =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0 ftype=0
log      =internal log           bsize=4096   blocks=521728, version=2
          =                       sectsz=512   sunit=0 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0

But they are there, and I am curious as to the developers/users 
thoughts.  Will they break things (nfs, CIFS/samba)?

Thanks!

Joe

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: question: are the metadata crc elements ready for general use?

[Dave Chinner]

On Tue, Apr 28, 2015 at 03:30:40PM -0400, Joe Landman wrote:
> I ask as I see this in the 3.2.2 mkfs.xfs command.  The default is
> to leave them off from what I can see on an mkfs.xfs.

Yes, and we've been discussing whether they should be made the
default in the upcoming 3.2.3 release or wait for 3.3.  IIRC, the
latest SLES release already defaults to using CRCs by default.

> root@unison:~/xfsprogs# mkfs.xfs /dev/sda
> meta-data=/dev/sda               isize=256    agcount=55,
> agsize=268435455 blks
>          =                       sectsz=512   attr=2, projid32bit=1
>          =                       crc=0        finobt=0
> data     =                       bsize=4096   blocks=14627632640, imaxpct=1
>          =                       sunit=0      swidth=0 blks
> naming   =version 2              bsize=4096   ascii-ci=0 ftype=0
> log      =internal log           bsize=4096   blocks=521728, version=2
>          =                       sectsz=512   sunit=0 blks, lazy-count=1
> realtime =none                   extsz=4096   blocks=0, rtextents=0
> 
> But they are there, and I am curious as to the developers/users
> thoughts.  Will they break things (nfs, CIFS/samba)?

I enable them on all my systems, and almost all of my testing is
done with "-m crc=1,finobt=1" options set. The don't affect
applications at all - CRCs are completely transparent to userspace
so applications don't even know the filesystem is using them.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs