Re: 2.6.12-rc6-mm1 & 2K lun testing

[Andrew Morton <akpm@osdl.org>]

Badari Pulavarty <pbadari@us.ibm.com> wrote:
>
> > It shouldn't be necessary to do both.  Either the patch or the tuning
> > should fix it.  Please confirm.
> > 
> > Also please determine whether the deep CFQ queue depth is a problem when
> > the VFS tuning/patching is in place.
> > 
> > IOW: let's work out which of these three areas needs to be addressed.
> > 
> 
> Andrew,
> 
> Sorry for not getting back earlier. I am running into weird problems.
> When running "dd" write tests to 2048 ext3 filesystems, just with your
> patch (no dirty ratio or CFS queue depth tuning), I see "buff" 
> increasing instead of "cache" and I see "bi" instead of "bo".
> Whats going on here ?

Beats me.  Are you sure you're not running a broken vmstat?

`buff' would increase if you were accidentally writing to /dev/sda1 rather
than /dev/sda1/some-filename, but I don't know why vmstat would be getting
confused over the direction of the I/O.

> 
> procs -----------memory---------- ---swap-- -----io---- --system-- 
> ----cpu----
>   r  b   swpd   free   buff  cache   si   so    bi    bo   in    cs us 
> sy id wa
> ..
>   2  0      4 6339920  42712  24884    0    0     0    19  413  1237 46 
>   6 48  0

You're wordwrapping...