GELI created on a GPT labelled partition doesn't work 2nd time around...

[Pawel Jakub Dawidek]

On Tue, Sep 30, 2014 at 09:40:55AM +0100, Karl Pielorz wrote:
> 
> 
> --On 28 September 2014 15:05 +0200 Pawel Jakub Dawidek <pjd at FreeBSD.org> 
> wrote:
> 
> > Could you provide the output of:
> >
> > 	# diskinfo -v /dev/gpt/abcdef.eli
> 
> Sure, after GPTing, geli init / geli attach - the running the above gives:
> 
> "
> /dev/gpt/abcdef.eli
>         4096            # sectorsize
>         750155304960    # mediasize in bytes (699G)
>         183143385       # mediasize in sectors
>         0               # stripesize
>         0               # stripeoffset
>         11400           # Cylinders according to firmware.
>         255             # Heads according to firmware.
>         63              # Sectors according to firmware.
>              WD-WCAPT0430257s0s0        # Disk ident.
> "
> 
> In the interim - I wrote a script which matches the drives serial numbers, 
> with keys - and does a 'geli attach' on the raw disk (e.g. /dev/da0) - this 
> results in da0.eli, which can then be GPT partitioned etc.
> 
> Trying to run anything 'against' /dev/gpt/abcdef.eli results in, e.g.
> 
> # gpart create -s gpt /dev/gpt/abcdef.eli
> gpart: provider: Device not configured
> 
> or,
> 
> dd if=/dev/gpt/abcdef.eli of=/dev/null bs=1k count=1000
> dd: /dev/gpt/abcdef.eli: Invalid argument

This fails because your gpt/abcdef.eli is configured with 4k sectors.
Try:

	dd if=/dev/gpt/abcdef.eli of=/dev/null bs=4k count=1000

> If I repeat the above steps without the initial GPT partition (i.e. against 
> /dev/da0) - I end up with '/dev/da0.eli' - which I can then GPT fine etc.

Note sure if GPT problem is due to 4k sector or because as jmg@
mentioned it checks for GPT in GPT configuration - a check we may be
able to relax.

-- 
Pawel Jakub Dawidek                       http://www.wheelsystems.com
FreeBSD committer                         http://www.FreeBSD.org
Am I Evil? Yes, I Am!                     http://mobter.com