Re: disk geometry issues when trying to set up encrypted partition

On Wed, Jun 16, 2010 at 05:45:40PM -0400, George Morgan wrote:

> Quoting Harry Palmer <[email protected]>:
> 
> >Hi there.
> >
> >I'm fairly new to openbsd and I'm hoping someone with better
> >understanding than me of how its disk handling works can help.
> >
> >Beginning my effort to encrypt a 300GB drive in a 64bit Ultrasparc,
> >I followed these initial steps:
> >
> >1. used disklabel to create a single slice "a" on the drive
> >
> >2. made a file system with newfs (is it necessary to have so many
> >   backup superblocks?)
> >
> >3. mounted sd2a on "/home/cy" and touched it with an empty file
> >     "/home/cy/cryptfile"
> >
> >4. zeroed out the file (and efectively the drive) with
> >     "dd if=/dev/zero of=/home/cy/cryptfile bs=512"
> >
> >
> >Here's the (eventual!) output of (4):
> >
> > /home/cy: write failed, file system is full
> > dd: /home/cy/cryptfile: No space left on device
> > 576520353+0 records in
> > 576520352+0 records out
> > 295178420224 bytes transferred in 19810.722 secs (14899932 bytes/sec)
> >
> >
> >
> >Now I have:
> >
> > # disklabel sd2a
> > # /dev/rsd2a:
> > type: SCSI
> > disk: SCSI disk
> > label: MAW3300NC
> > flags: vendor
> > bytes/sector: 512
> > sectors/track: 930
> > tracks/cylinder: 8
> > sectors/cylinder: 7440
> > cylinders: 13217
> > total sectors: 585937500
> > rpm: 10025
> > interleave: 1
> > boundstart: 0
> > boundend: 585937500
> > drivedata: 0
> >
> > 16 partitions:
> > #                size           offset  fstype [fsize bsize  cpg]
> >   a:        585937200                0  4.2BSD   2048 16384    1
> >   c:        585937500                0  unused
> >
> >
> >and:
> >
> > # ls -l /home/cy
> > total 576661216
> > -rw-r--r--  1 root  wheel  295178420224 Jun 16 03:39 cryptfile
> >
> >
> >and:
> >
> > # df -h
> > Filesystem     Size    Used   Avail Capacity  Mounted on
> > /dev/sd0a     1007M   44.8M    912M     5%    /
> > /dev/sd0k      247G    2.0K    235G     0%    /home
> > /dev/sd0d      3.9G    6.0K    3.7G     0%    /tmp
> > /dev/sd0f      2.0G    559M    1.3G    29%    /usr
> > /dev/sd0g     1007M    162M    795M    17%    /usr/X11R6
> > /dev/sd0h      5.9G    212K    5.6G     0%    /usr/local
> > /dev/sd0j      2.0G    2.0K    1.9G     0%    /usr/obj
> > /dev/sd0i      2.0G    2.0K    1.9G     0%    /usr/src
> > /dev/sd0e      7.9G    7.7M    7.5G     0%    /var
> > /dev/sd2a      275G    275G  -13.7G   105%    /home/cy
> >
> >
> >
> >I have no understanding of this. I've never seen a df output
> >that tells me I'm using 13GB more space than the drive is
> >capable of holding.
> >
> >I ask here because there's obviously potential for me to lose
> >data somewhere down the line. I'll be grateful if anyone can
> >explain where I've gone wrong.
> 
> I've seen the greater than 100% full on a UFS? filesystem before when
> you exceed the size of the filesystem.  There is space in the
> filesystem for "lost+found" and all those superblocks? you were
> complaining about that can get overwritten if you write too much to a
> partition.

Spoace for superblocks and other metadata is subtracted from available
blocks. lost+found is an ordinary directory. 

> 
> So setting up your "dd" to actually stop before you overfill the
> filesystem is what you need to do. (using bs=# count=# ... info you
> can get before you start initializing your file with the df command
> without the "-k or -h" to get number of blocks and block size)
> 
> I'm sure the fine people on these lists will correct me if I'm wrong
> in my assumptions...  :-)

You are wrong, there;s no such thing as overfilling a filesystem. It's
just the 5% reserved for root. An ordinary user runs out earlier. It's
in the FAQ. 


	-Otto
> 
> George Morgan