On Thu, 17 Feb 2005 16:08:57 -0800, Carl Lowenstein
<[EMAIL PROTECTED]> wrote:
> On Thu, 17 Feb 2005 15:29:54 -0800, Gregory K. Ruiz-Ade
> <[EMAIL PROTECTED]> wrote:
> > On Thursday 17 February 2005 03:08 pm, Carl Lowenstein wrote:
> > > I painfully went through the process to rename VolGroup00-LogVol00 to
> > > VG00-LV00, etc., just to make the names shorter, but this did not
> > > help.
> > >
> > > So shell scripts that parse df(1) output by looking for lines that
> > > begin with /dev no longer work as intended. The data that belongs to
> > > the /dev/xxx name is now on the following line, if it's a partition
> > > managed by LVM.
> >
> > I'm guessing your shorter names are still beyond the threshold that df uses
> > to keep the data on a single line. df seems to do that only with devices
> > with long names...
> >
> > My other question is why LVM devices were showing up
> > as /dev/mapper/<vg>-<lv>, unless you specifically created them that way.
> > I'm accustomed to the /dev/<volumegroup>/<logicalvolume> naming scheme,
> > which still lets you be nice and short:
> >
> > /dev/SysVol/boot
> > /dev/SysVol/root
> > /dev/SysVol/swap
> > /dev/SysVol/tmp
> >
> > etc.
> >
> > for a time I used HPUX's convention of just numbers, i.e.,
> >
> > /dev/vg00/lv01
> >
> > but found that naming things made it easier for me.
> >
> > And yes, I realize that I have been mostly unhelpful.
>
> The names are set up as /dev/VG00/LV00 etc, with lines in /etc/fstab such as:
> /dev/VG00/LV06 /data ext3 defaults 1 2
> It is df(1) that reformats the information to look like:
> /dev/mapper/VG00-LV06
>
> A quick look through the source for coreutils-5.2.0 which I have here
> online makes me believe that this special treatement of logical
> volumes must have started in coreutils-5.2.1. Current version is
> 5.2.1-31.
>
> I'm heading off to look for those sources. Just now I tried df from
> coreutils-5.0, and it does the same two-line thing. So I must have
> missed it in my quick source scan. Bah.
Use the source, Luke. Then find that the option is somewhat poorly
explained in the man.page. Option "-P" use POSIX output format
makes the 2-line output back into the single line that we all know and
expect from previous experience.
carl
--
carl lowenstein marine physical lab u.c. san diego
[EMAIL PROTECTED]
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
