[PATCH] mkfs: add missing noalign suboption information

[PATCH] mkfs: add missing noalign suboption information

[Carlos Maiolino]

Adds to the mkfs man page, information about noalign mkfs option which is
missing until now.

Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
---
 man/man8/mkfs.xfs.8 | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/man/man8/mkfs.xfs.8 b/man/man8/mkfs.xfs.8
index 04c19b4..2435814 100644
--- a/man/man8/mkfs.xfs.8
+++ b/man/man8/mkfs.xfs.8
@@ -244,6 +244,12 @@ will automatically query the logical volume for appropriate
 and
 .B swidth
 values.
+.TP
+.BI noalign
+This option is used to bypass data alignment configuration. In rare
+circunstances, data alignment might decrease performance, for such cases,
+use this option to avoid data alignment, this will set stripe unit and
+stripe width to 0.
 .RE
 .TP
 .B \-f
-- 
1.8.1.4

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

Re: [PATCH] mkfs: add missing noalign suboption information

[Eric Sandeen]

On 7/8/13 6:45 PM, Carlos Maiolino wrote:
> Adds to the mkfs man page, information about noalign mkfs option which is
> missing until now.
> 
> Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
> ---
>  man/man8/mkfs.xfs.8 | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/man/man8/mkfs.xfs.8 b/man/man8/mkfs.xfs.8
> index 04c19b4..2435814 100644
> --- a/man/man8/mkfs.xfs.8
> +++ b/man/man8/mkfs.xfs.8
> @@ -244,6 +244,12 @@ will automatically query the logical volume for appropriate
>  and
>  .B swidth
>  values.
> +.TP
> +.BI noalign
> +This option is used to bypass data alignment configuration. In rare
> +circunstances, data alignment might decrease performance, for such cases,
> +use this option to avoid data alignment, this will set stripe unit and
> +stripe width to 0.

Thanks for this, Carlos.

I think it might be better to just state plainly that it overrides automatic
stripe unit detection, and leave the tuning / performance parts out.
Most of the manpage today is more "what" than "why."

So maybe simply:

+This option disables automatic stripe geometry detection, and creates
+the filesystem without stripe geometry even if the underlying storage
+device reports this information.

FWIW, there is a 'noalign' option for the realtime subvolume
as well, and that should probably get documented too.

-Eric

>  .RE
>  .TP
>  .B \-f
> 

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

Re: [PATCH] mkfs: add missing noalign suboption information

[Carlos Maiolino]

On Mon, Jul 08, 2013 at 07:01:37PM -0500, Eric Sandeen wrote:
> On 7/8/13 6:45 PM, Carlos Maiolino wrote:
> > Adds to the mkfs man page, information about noalign mkfs option which is
> > missing until now.
> > 
> > Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
> > ---
> >  man/man8/mkfs.xfs.8 | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/man/man8/mkfs.xfs.8 b/man/man8/mkfs.xfs.8
> > index 04c19b4..2435814 100644
> > --- a/man/man8/mkfs.xfs.8
> > +++ b/man/man8/mkfs.xfs.8
> > @@ -244,6 +244,12 @@ will automatically query the logical volume for appropriate
> >  and
> >  .B swidth
> >  values.
> > +.TP
> > +.BI noalign
> > +This option is used to bypass data alignment configuration. In rare
> > +circunstances, data alignment might decrease performance, for such cases,
> > +use this option to avoid data alignment, this will set stripe unit and
> > +stripe width to 0.
> 
> Thanks for this, Carlos.
> 
> I think it might be better to just state plainly that it overrides automatic
> stripe unit detection, and leave the tuning / performance parts out.
> Most of the manpage today is more "what" than "why."
> 
> So maybe simply:
> 
> +This option disables automatic stripe geometry detection, and creates
> +the filesystem without stripe geometry even if the underlying storage
> +device reports this information.
> 
> FWIW, there is a 'noalign' option for the realtime subvolume
> as well, and that should probably get documented too.
> 

Agreed, going to change and re-send it.

> -Eric
> 
> >  .RE
> >  .TP
> >  .B \-f
> > 
> 
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs

-- 
Carlos

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