Subject: RE: NetBSD Hang after disklabel, xsm crash
To: None <harald@bitcon.no>
From: None <Havard.Eidnes@runit.sintef.no>
List: port-i386
Date: 07/28/1998 15:55:26
> > > For the 3,2GB I did the same thing. After the first "newfs
> > > wd1c", I got a message telling me that I was wasting 8064
> > > sectors of the disk. I added 8064 to the number 6281856
> > > which was the size I had entered for my C-partition on this
> > > disk, and tried to do a new disklabel.
> >
> > Did you alter your geometry information as well? Perhaps a
> > sector "outside" the disklabel's geometry causes internal
> > problems in the filesystem code.
>
> I only extended the C partition size-information by 8064 sectors.

Correct me if I'm wrong, but I *think* the message from newfs
does not indicate that the wasted space is *outside* the given
partition, but *inside* the partition.  My guess is that you got
the warning printed by these two printf() statements in mkfs.c:

printf("Warning: inode blocks/cyl group (%d) >=3D data blocks (%d) in last\n",
printf("    cylinder group. This implies %d sector(s) cannot be allocated.\n",

This is entirely normal, and just indicates that the last
cylinder group on the disk is not large enough to even hold the
inode blocks for that cylinder group, so the space is "wasted"
and cannot be used by the file system.  You can usually get
around this if you are willing to tweak the newfs paramters.

- H=E5vard