[Info-vax] Problems restoring image-backup

[David Froble]

Stephen Hoffman wrote:
> On 2016-06-28 19:35:44 +0000, David Froble said:
> 
>> My understanding is in an /IMAGE restore that all volume info comes 
>> from the save set.
> 
> Nope.  Some contents change.
> 
>> Perhaps this is not correct?
> 
> Ayup.   It's not correct.
> 
>> If not, then I'd consider it a bug.
> 
> You can consider it a bug, but it's how the Itanium boot environment 
> necessarily works,

And that's fine.  A switch something like /INSTALL_ITANIC_JUNK could then 
deviate from a pure image copy.

Note, I'm well aware that it's a file structured copy, not a true physical image.

With a save set created with /IMAGE there is already some disk information used 
to init the disk.  Sure seems simple to me to also have some hackery for knowing 
where to point the bootstrap.  Would still all come from the save set.

> particularly with the hack-around that exists here 
> for the lack of partitioning support within OpenVMS I64, as the EFI 
> console requires the disk to be partitioned.   But it's also how the VAX 
> and Alpha environments necessarily work with BACKUP, too.   The console 
> bootstrap-related data structures involved here contain pointers to 
> sectors.  With BACKUP /IMAGE, those sectors inherently change, which 
> means that BACKUP can't slam in the old contents and boot.

The address of the sectors will change, but it's still the same files, and from 
the save set.

>  On all three 
> architectures, the boot structures are mapped to files.  Those file 
> contents cannot be completely restored unmodified; the sector pointers 
> have to change.

Agreed.