On Tue, 1 Feb 2000, Dirk Hoppe wrote:

> I need some help in reading the disklabel and the output of the df comman=
d.
> I made partitions with FWB HDT in the first part of the HD: for sd1a 50 M=
B
> (root), for sd1b 160 MB (swap), for sd1g 700 MB (/usr) and for sd1e 150 M=
B
> (/home). Partition sd1f isn=B4t used yet. I formatted them with MKfs 1.47=
 and
> installed the system. The output of the disklabel and df command are as
> followed. I was surprised about the partition size in disklabel. For exam=
ple
> a: 102400, shouldn=B4t it be 51200 (50 MB)?

Dirk,
=09here's part of a a message that I sent just yesterday.  It may
answer your questions.=20


=3D=3D>>

On Mon, 31 Jan 2000, Christopher P. Gill wrote:

> On Mon, 31 Jan 2000, Merideth Johnston wrote:
>=20
> > the bad news is, the other partition sd0d is, well, wrong.  I know it's
> > wrong because it only has one file on it, and that is an empty director=
y,
> > yet df tells me it is 75% full (what _does_ it think is _in_ there?), a=
nd
> > more than twice as large as it's supposed to be!?!?  This makes me nerv=
ous.
> > I'd like to just reformat the thing and start over with it.  I've alrea=
dy
> > re-made the directory in question elsewhere, and wondered if I could do=
 the
> > reformatting of that partition without using the little devil on it aga=
in,
> > as that did seem to have some problems with the large drive.  (I'm hopi=
ng
> > that is the cause of sd0d's pecular problem)  Can I instead just use
> > something from within NetBSD to reformat that partition?  (wiping out t=
he
> > erroneous size thing, and all else - whatever that else is - with it)  =
The
> > size of it appears correct when listed with disklabel, so it seems to m=
e it
> > would just be the contents, the file structure, that needs to be redone=
=2E
>=20
> Meredith,
> =09there is a known bug (well, I found out about it the same way
> that you did) having to do with df reporting partition sizes incorrectly
> on partitions subsequent to the first on a disk.  It kind of adds the
> values from the first partition to the next and so on, and reports
> values that are too high.

<<=3D=3D


Try checking with 'du', and see if you've really used up the space that
'df' reports.


/*=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
"Don't die wondering..."                http://www.cldc.howard.edu/~cpg
                                              email: cpg@scs.howard.edu
chris out-              Christopher P. Gill
  peace.        C.L.D.C. Senior System Operator (Ret.)
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D*/